This action might not be possible to undo. Are you sure you want to continue?
CONTENTS
Preface 1 Terror, tragedy and bad vibrations 1.1 Introduction . . . . . . . . . . 1.2 The Tower of Terror . . . . . 1.3 Into thin air . . . . . . . . . . 1.4 Music and the bridge . . . . . 1.5 Discussion . . . . . . . . . . . 1.6 Rules of calculation . . . . . . Functions 2.1 Rules of calculation . . . 2.2 Intervals on the real line . 2.3 Graphs of functions . . . 2.4 Examples of functions .
v 1 1 1 4 7 8 9 11 11 15 17 20 27 27 30 41 41 48 53 57 57 62 65 68 70 75 75 81 84 90
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
2
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
3
Continuity and smoothness 3.1 Smooth functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Differentiation 4.1 The derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Rules for differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Velocity, acceleration and rates of change . . . . . . . . . . . . . . . . . . . . . . . Falling bodies 5.1 The Tower of Terror . . . . . . . . 5.2 Solving differential equations . . . . 5.3 General remarks . . . . . . . . . . . 5.4 Increasing and decreasing functions 5.5 Extreme values . . . . . . . . . . . Series and the exponential function 6.1 The air pressure problem . . . . 6.2 Inﬁnite series . . . . . . . . . . 6.3 Convergence of series . . . . . . 6.4 Radius of convergence . . . . .
4
5
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
6
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
ii
CONTENTS
6.5 6.6 6.7 6.8 7
Differentiation of power series . . . . The chain rule . . . . . . . . . . . . . Properties of the exponential function Solution of the air pressure problem .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. 93 . 96 . 99 . 102 109 109 111 114 119 122 125 127 127 134 140 143 145 153 167 167 168 170 171 174 176 182 187 188 197 200 205 214 218 221
Trigonometric functions 7.1 Vibrating strings and cables . . . . . . . . . . 7.2 Trigonometric functions . . . . . . . . . . . 7.3 More on the sine and cosine functions . . . . 7.4 Triangles, circles and the number . . . . . . 7.5 Exact values of the sine and cosine functions . 7.6 Other trigonometric functions . . . . . . . . . Oscillation problems 8.1 Second order linear differential equations 8.2 Complex numbers . . . . . . . . . . . . . 8.3 Complex series . . . . . . . . . . . . . . 8.4 Complex roots of the auxiliary equation . 8.5 Simple harmonic motion and damping . . 8.6 Forced oscillations . . . . . . . . . . . . Integration 9.1 Another problem on the Tower of Terror . 9.2 More on air pressure . . . . . . . . . . . 9.3 Integrals and primitive functions . . . . . 9.4 Areas under curves . . . . . . . . . . . . 9.5 Area functions . . . . . . . . . . . . . . . 9.6 Integration . . . . . . . . . . . . . . . . . 9.7 Evaluation of integrals . . . . . . . . . . 9.8 The fundamental theorem of the calculus . 9.9 The logarithm function . . . . . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
8
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
9
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
10 Inverse functions 10.1 The existence of inverses . . . . . . . . 10.2 Calculating function values for inverses 10.3 The oscillation problem again . . . . . 10.4 Inverse trigonometric functions . . . . . 10.5 Other inverse trigonometric functions .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
11 Hyperbolic functions 225 11.1 Hyperbolic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 11.2 Properties of the hyperbolic functions . . . . . . . . . . . . . . . . . . . . . . . . . 227 11.3 Inverse hyperbolic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
CONTENTS
iii 235 235 237 239 241 243 247 249
12 Methods of integration 12.1 Introduction . . . . . . . . . . . . . . 12.2 Calculation of deﬁnite integrals . . . . 12.3 Integration by substitution . . . . . . 12.4 Integration by parts . . . . . . . . . . 12.5 The method of partial fractions . . . . 12.6 Integrals with a quadratic denominator 12.7 Concluding remarks . . . . . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
13 A nonlinear differential equation 251 13.1 The energy equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 13.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 Answers Index 261 281
PREFACE
If I have seen further it is by standing on the shoulders of Giants. Sir Isaac Newton, 1675. This book presents an innovative treatment of single variable calculus designed as an introductory mathematics textbook for engineering and science students. The subject material is developed by modelling physical problems, some of which would normally be encountered by students as experiments in a ﬁrst year physics course. The solutions of these problems provide a means of introducing mathematical concepts as they are needed. The book presents all of the material from a traditional ﬁrst year calculus course, but it will appear for different purposes and in a different order from standard treatments. The rationale of the book is that the mathematics should be introduced in a context tailored to the needs of the audience. Each mathematical concept is introduced only when it is needed to solve a particular practical problem, so at all stages, the student should be able to connect the mathematical concept with a particular physical idea or problem. For various reasons, notions such as relevance or just in time mathematics are common catchcries. We have responded to these in a way which maintains the professional integrity of the courses we teach. The book begins with a collection of problems. A discussion of these problems leads to the idea of a function, which in the ﬁrst instance will be regarded as a rule for numerical calculation. In some cases, real or hypothetical results will be presented, from which the function can be deduced. Part of the purpose of the book is to assist students in learning how to deﬁne the rules for calculating functions and to understand why such rules are needed. The most common way of expressing a rule is by means of an algebraic formula and this is the way in which most students ﬁrst encounter functions. Unfortunately, many of them are unable to progress beyond the functions as formulas concept. Our stance in this book is that functions are rules for numerical calculation and so must be presented in a form which allows function values to be calculated in decimal form to an arbitrary degree of accuracy. For this reason, trigonometric functions ﬁrst appear as power series solutions to differential equations, rather than through the common deﬁnitions in terms of triangles. The latter deﬁnitions may be intuitively simpler, but they are of little use in calculating function values or preparing the student for later work. We begin with simple functions deﬁned by algebraic formulas and move on to functions deﬁned by power series and integrals. As we progress through the book, different physical problems give rise to various functions and if the calculation of function values requires the numerical evaluation of an integral, then this simply has to be accepted as an inconvenient but unavoidable property of the problem. We would like students to appreciate the fact that some problems, such as the nonlinear pendulum, require sophisticated mathematical methods for their analysis and difﬁcult mathematics is unavoidable if we wish to solve the problem. It is not introduced simply to provide an
vi PREFACE intellectual challenge or to ﬁlter out the weaker students. forced damped oscillations and the nonlinear pendulum. there seems to be little need for science and engineering students to spend time evaluating anything but the simplest of integrals by hand. We discuss continuity and differentiation in terms of convergence of sequences. Various ad hoc procedures which might have sufﬁced for some of the earlier problems are no longer useful. but not all of them need formal proof. If limits are treated with the full rigour of the . The use of Mathematica makes plotting of elliptic functions and ﬁnding their values no more difﬁcult than is the case with any of the common functions. However. In each case the solution is a function which relates two variables. Methods of integration are introduced as a practical alternative to numerical methods for evaluating integrals if a primitive function can be found. Students are quite capable of using the results on termbyterm differentiation of a power series for instance. Our attitude to proofs and rigour is that we believe that all results should be correctly stated. We do not go very deeply into this topic. Tim Langtry and Graeme Cohen read the text of the preliminary edition of this book with meticulous attention and made numerous suggestions. Modelling the problems leads to differential equations for the desired functions and in solving these equations we discuss power series. In discussing oscillation we have to consider the case where the auxiliary equation may have nonreal roots and it is at this point that we introduce complex numbers. Other useful suggestions. we state the result and move on. By this means we can take the applications in this book beyond the artiﬁcial examples often seen in standard texts. With the advent of symbolic manipulation packages such as Mathematica. Otherwise. A We would like to thank Tim Langtry for help with LTEX. Our list of physical problems includes the vertical motion of a projectile. then they are too difﬁcult for the average beginning student. ¢ ¡ . then we provide it. the motion of a body in simple harmonic motion. Closer analysis of these intuitive ideas leads to the concepts of continuity and differentiability. the variation of atmospheric pressure with height. We also need to know whether a function deﬁned by an integral is new or whether it is a known elementary function in another form. If we feel that a proof is accessible and that there is something useful to be learned from the proof. We think that this is intuitively more accessible than the usual approach of considering limits of functions. we think that it is important to emphasise that a power series can be differentiated in this way only within the interior of its interval of convergence. An appeal to the student’s physical intuition suggests that the graphs of these functions should have certain properties. to deﬁne functions by integrals and to deﬁne inverse functions. contributions and corrections came from Mary Coupland and Leigh Wood.approach. while a nonrigorous treatment simply leads to confusion. radius of convergence and termbyterm differentiation. The book concludes with a capstone discussion of the nonlinear equation of motion of the simple pendulum. Most of all. underdamped and overdamped oscillations. Not all differential equations are amenable to a solution by power series and integration is developed as a method to deal with these cases. even if they have not seen the proof. we do not believe that students should be presented with handwaving arguments masquerading as proofs. comments and corrections. Along the way it is necessary to use the chain rule. Our purpose here is to demonstrate the fact that there are physical problems which absolutely need the mathematics developed in this book. The remainder of this preface summarises the content of this book.
It comes momentarily to rest and then free falls for about ﬁve seconds to again reach a speed of almost . some people do have to understand the underlying principles of such devices in order to invent them in the ﬁrst place.2 THE TOWER OF TERROR Sixteen people are strapped into seats in a six tonne carriage at rest on a horizontal metal track.CHAPTER 1 TERROR. By exploring each of these in increasing detail we will ﬁnd it necessary to introduce a large number of mathematical techniques in order to obtain solutions to the problems. Part of the purpose of this book is to demonstrate the way that mathematics pervades many aspects of our lives. These problems will be used as the basis for introducing new mathematical ideas and in later chapters we will apply these ideas to the solution of other problems. However. but this use is not always evident. as in many other areas of knowledge. someone must design the systems which enable these activities to be carried out. but once again. but when would we ever use algebra or calculus? In mathematics. but is applicable to many other situations. The power is switched on and in six seconds they are travelling at 160 km/hr. simple arithmetic is needed to live a normal life in developed societies. scientists and economists to manage them. People do not have to understand how a car. we shall make use of three easily understood and obviously relevant problems. Most people do not need to know how to organise the Olympic Games. The carriage traverses a short curved track and then hurtles vertically upwards to reach the height of a 38 storey building. TRAGEDY AND BAD VIBRATIONS 1. the disastrous consequences that occurred when an aircraft cargo door ﬂew open in midair and an unexpected noise pollution problem on a new bridge. to improve their design or to repair them. a computer or a mobile phone works in order to make use of them.1 INTRODUCTION Mathematics is almost universally regarded as a useful subject. Certainly. we can often get by with a less than complete understanding of the processes. we shall ﬁnd that it is not limited to the original problems. schedule baggage handlers for an international airline or analyse trafﬁc ﬂow in a communications network. 1. but the truth of the matter is that mathematics beyond the middle levels of high school is almost never used by the ordinary person. we will consider three problems: an amusement park ride known as the Tower of Terror. social and ﬁnancial systems used by our modern society all rely on mathematics to a greater or lesser extent and we need skilled people such as engineers. As we become more familiar with the mathematics we develop. To do this. In this chapter. The complex technical. Mathematics is widely used.
TRAGEDY AND BAD VIBRATIONS Figure 1.1: The Tower of Terror .2 TERROR.
The carriage is accelerated along a horizontal track from the starting point . many rides have restrictions on who can take the ride and there are often warning notices about the danger of taking the ride for people with various medical problems. with the resulting shower of fastmoving debris over the park.2: The Tower of Terror (Schematic) 160 km/hr. Second. the riders experience the sensation of weightlessness for ﬁve or six seconds. It hurtles back around the curve to the horizontal track where powerful brakes bring it to rest back at the start. Let’s look at some aspects of the ride in the Tower of Terror illustrated in the schematic diagram in Figure 1. with the consequent increase in cost and difﬁculty of construction. In fact. is the need to ensure that patrons will be able to physically withstand the forces to which they will be subjected. Australia. The motion is then reversed as the carriage free falls back to . The time depends on the speed at which the carriage is travelling when it reaches on the outward journey and the higher this speed the longer the horizontal portion of the track must be and the more power is required to ¦ £ ¥ £ ¦ ¥ ¥ ¤ ¤ ¥ ¥ £ ¤ ¥ ¦ . If the time is too short then the ride would be pointless. would be disastrous. This is the time during which the riders experience weightlessness during free fall. The carriage then goes round the curved section of the track to reach the horizontal portion of the track. the brakes are applied at and the carriage comes to a stop at . From the speed decreases under the inﬂuence of gravity until it comes momentarily to rest at . The most important feature of the ride is perhaps the time taken for the carriage to travel from back to .1). The whole event takes about 25 seconds (Figure 1. and equally important. the higher the tower must be. faster and more exciting.2. it is travelling at 160 km/hr and it then travels around a curved portion of track until its motion has become vertical by the point . When it reaches after about six seconds.THE TOWER OF TERROR 3 Figure 1. 115 metres above the ground or the height of a 38 storey building. Parks like this have become common around the world with the best known being Disneyland in the United States. One of the main features of the parks are the rides which are offered and as a result of competition between parks and the need to continually change the rides. During this portion of the ride. The longer the time however. These trends have resulted in the development of a specialised industry to develop and test the rides which the parks offer. a large amusement park on the Gold Coast in Queensland. There are two aspects to this. Such failure. This hairraising journey takes place every few minutes at Dreamworld. they have become larger. The ride just described is aptly named the Tower of Terror. First the construction must ensure that the equipment will not collapse under the strains imposed on it.
Almost all of Aristotle’s methods for analysing motion have turned out to be wrong. About half an hour later. extensive enquiries were conducted to ﬁnd the cause. This is a modern version of the problem of the motion of falling bodies. Aristotle’s ideas about motion went almost unchallenged for many centuries and it was not until the 14th century that a new approach to many of the problems of physics began to emerge. Other Greek thinkers were also producing such ambitious systems of ideas. it required great insight on his part to do this. there was an outrush of air from the cabin with such force that nine passengers were sucked out and never seen again.3 INTO THIN AIR At 1. The development of the law of falling bodies began in Greece about 300 BC. who in 1638 published his Dialogues Concerning the Two New Sciences in which he presented his ideas on the principles of mechanics. the US National Transportation Safety Board concluded that a faulty switch in the door control system had caused it to open. The aircraft turned back to Honolulu and.500 years.m. The strong current of air ¦ ¥ ¥ ¦ . but no bodies were found. The design of the ride is thus a compromise between the time taken for the descent. along with 7 square metres of the fuselage. With nobody to show him how to solve the problem. After inspecting the door and considering all other evidence. but Aristotle was the only one to place much importance on the analysis of motion as we would now understand the word. the cost of construction and the power consumed on each ride. A coast guard search under the ﬂight path located 57 pieces of material from the aircraft. the height of and the time taken to .33 a. but the physical cause of the subsequent events was the explosive venting of air from the aircraft. From observation of everyday motions. As a result of this event. We shall derive the law from a hypothetical set of experimental results to illustrate the way in which mathematical methods develop. a to travel from problem which has been discussed for about 2. landed at 2. the forward cargo door opened without warning. The Board made recommendations about procedures which would prevent such accidents in future and stated that proper corrective action after a similar cargo door incident in 1987 could have prevented the tragedy. the Greek philosopher Aristotle put forward a collection of results about the motion of falling bodies as part of a very large system of ideas that was intended to explain the whole of reality as it was then known. The ﬁrst task is to ﬁnd the relation between the speed at . but he was nevertheless the ﬁrst to introduce the idea that motion could be analysed in numerical terms. The two forward toilet compartments were displaced by 30 cm. 1.34 a. and was torn off. At that time Greece was the intellectual centre of the western world and there were already two hundred years of scientiﬁc and philosophical thought to build upon. TRAGEDY AND BAD VIBRATIONS accelerate the carriage on each ride.4 TERROR. The cargo door was located and recovered in two separate pieces by a United States Navy submarine in October 1990. with considerable difﬁculty. The event which triggered the accident was the opening of the cargo door. He was the ﬁrst person to give an accurate explanation of the motion of falling bodies in more or less modern terms. on 24 February 1989 ﬂight UAL 811 left Honolulu International Airport bound for Auckland and Sydney with 337 passengers and 18 crew on board. Perhaps the ﬁrst real physicist in modern terms was Galileo. when the aircraft was over the ocean 138 km south of the airport and climbing through 6700 m.m. everybody could see that the problem was an easy one to solve and it is now a routine secondary school exercise. Six tyres blew out during the landing and the brakes seized. All ten emergency slides were used to evacuate the passengers and crew and this was achieved with only a few minor injuries. As with all aircraft accidents. Two of the engines and parts of the starboard wing were damaged by objects emerging from the aircraft and the engines had to be shut down. But once Galileo had done the hard work.
Firstly. In addition. In fact. the mercury in the tube always dropped to a level of about Pressure 76 cm above the mercury surface in the open vessel. it is difﬁcult to inhale sufﬁcient air by normal breathing. as occurred in the accident with ﬂight 811.3: Flight UAL811 was apparent to all on board. As shown in Figure 1. Secondly. He closed the open end with a ﬁnger and then inverted the tube with Mercury the open end in a vessel containing mercury. The physical mechanism which is at the heart of the events described above is also involved in a much less dramatic phenomenon and it was in this other situation that the explanation of the mechanism ﬁrst emerged historically.6 gm/cm and so the weight of a column of mercury of unit area and 76 cm high is 1030 gm.4. as a result. water barometers had been constructed a few years before Torricelli and it had been found the maximum height of the column of water was 10. § § . In 1643 Evangelista Torricelli. The density of mercury is 13. the air is not sucked out.4.4 m. the phenomenon does not occur at low altitudes. the same breathing difﬁculties are experienced on high mountains when no motion at all is taking place. constructed the ﬁrst modern barometer. and that. After the initial outrush of air. given Figure 1. A ﬁrst attempt to explain this event might be that the speed of the aircraft through the still air outside caused the air inside to be sucked out. the weight of the liquid in the tube is exactly balanced by the weight of the atmosphere pushing down on the liquid surface. but passengers found it difﬁcult to breathe. There are several reasons why this is not convincing. water or mercury. This is shown schematically in Vacuum Figure 1. there will be a ﬂow of air into the region where the air is thin.4: Barometer that the density of water is 1gm/cm . if air at normal sea level pressure is brought in contact with the thin upper level air. The simplest way to describe these experiments is in terms of the pressure exerted by the column of ﬂuid. If a window is opened in a fast moving car or a low ﬂying light aircraft. whether air.4 m.INTO THIN AIR 5 Figure 1. This weight is almost identical to the weight of a column of water of unit area and height 10. It appears that the atmosphere becomes thinner in some way as height increases. the situation in the aircraft stabilised. Torricelli took a long glass tube and ﬁlled it with mercury. When the ﬁnger was Atmospheric released. a friend and follower of Galileo.
a set of experimental results will be used to obtain at least an approximate form for the required law.6 TERROR. the internal pressure forced air in the aircraft out of the opening until the internal and external pressures were equal. mountaineering and so on. As aircraft cabins are pressurised. As with the falling body problem. The external pressure would have been less than half this value.) Suppose the pool were ﬁlled with seawater ( kg/m ). EXERCISES 1. It is known from physical principles that the pressure exerted at a depth in an incompressible ﬂuid (such as water) is given by N/m . Examples include dentists’ chairs. since we need this pressure to force air into our lungs. but the variation of atmospheric pressure with height does not follow a simple rule. The column of mercury in a barometer is 75 cm high. where is the density of the ﬂuid and is the acceleration due to gravity—approximately m/s . Compute the air pressure in kg/m . 5 m wide and 2 m deep. Obtaining the law of pressure variation from this set of experimental results will be more difﬁcult than in the case of a falling body. § % ' % " ¨ " ! ¨ © § % % % " . What force is exerted on the bottom of the pool by the weight of the water? (You kg/m .3 1. TRAGEDY AND BAD VIBRATIONS Height Pressure 10 000 m 264 mb Height Pressure 5000 m 540 mb Height Pressure 0m 1013 mb Earth Figure 1.5: Variation of pressure with height It was soon found that atmospheric pressure is not constant even at sea level and that the small variations in pressure are related to changes in the weather. A swimming pool is 8 m long. There are many common devices which utilise ﬂuid pressure. It was also found that the pressure decreases with height above sea level and this is to be expected since the mass of the column of air decreases with height (Figure 1. We can now give a partial explanation of the events of Flight 811.5). These results will be the average value for the pressure at various heights in the atmosphere. at which time the situation stabilised. What other examples can you think of? 2. aircraft design. What may take force is now exerted on the bottom of the pool by the weight of the water? 3. car lifts and hydraulic brakes. When the cargo door burst open. the pressure inside the cabin would have been approximately that of normal ground level pressure. The difﬁculty in breathing would have been due to the reduced pressure. It is essential to have some model for the variation of pressure with height because of the needs of weather forecasting.
4 MUSIC AND THE BRIDGE Almost everyone enjoys a quiet night at home. much to the annoyance of local residents. When the piston is moving down. As with the previous two problems.MUSIC AND THE BRIDGE 7 4. loud parties. As originally designed. New forms of noise pollution are continually arising and some of these are quite unexpected. An example of this is one of the most famous bridge collapses in history. then the system will vibrate strongly and in some systems catastrophic results can follow if the vibrations are not damped out. The main span of this concrete bridge is 345 m long and 32. . A pump is a device which occurs in many situations—pumping fuel in automobiles. Engineers working on the bridge had to ﬁnd a way to damp the vibrations and thereby reduce or eliminate the noise. Explain also why the maximum height to which water can be raised with such a pump is about 10. the valve in the cylinder closes and Water the valves in the piston open. It is the cables which created an unexpected problem. A moderately strong wind started the bridge vibrating with its natural frequency. The results were spectacular. In these instruments a metal string is stretched between two supports and when the string is displaced by plucking or rubbing it begins to vibrate and emit a musical note. they were enclosed in polyurethane coverings. this problem is a modern version of one that has been in existence for many centuries. the best known of these being the Sydney Harbour Bridge. trafﬁc din or sporting events. Whether it is aircraft noise. When the piston moves up.4 m. In the case of the bridge. originally known as the Glebe Island Bridge (Figure 1. A simple type of water pump is Piston shown in the diagram on the left. The cylinder has a valve at Valves its lower end and there are valves on the piston.6).2 m wide. The deck is supported by two planes of stay cables attached to two 120 m high reinforced concrete towers. the cables began to vibrate and bang against the coverings. The newest bridge is the A$170m Anzac Bridge. The analysis of the vibrations is a complex problem which can be approached in stages. A cylinder containing a Upper piston is lowered into a tank. but in the modern world there is less and less opportunity for this simple pleasure. The form in which it principally arose in the past was in relation to the sounds made by stringed instruments such as violins and guitars. the valve in the cylinder opens and the valves in the piston Lower close. When the wind was at a certain speed from the southeast. beginning with the simplest possible type of model. This occurred on 7 November 1940. when the Tacoma Narrows Bridge in the United States had only been open for a few months. Sydney contains a large number of bridges. pumping water from a tank or borehole or pumping gas in air conditioners or refrigerators. News movies show the entire bridge oscillating wildly in a wavelike motion before it was ﬁnally wrenched apart. Any vibrating system has a natural frequency at which it will vibrate if set in motion. Based on the discussion on air pressure given in the Valve text. there are many forms of noise pollution which cause annoyance or disruption. explain how such a system can be used to pump water from the tank. The resulting noise could be heard several kilometres away from the bridge. 1. the aim is to prevent the vibrations or else to damp them out as quickly as possible when they begin. The frequency of the note is the same as the frequency of the vibrations of the string and so the problem becomes one of relating the frequency of vibration to the properties of the string. If a force is applied to the system which tries to make it vibrate at this frequency.
the difﬁculty experienced by passengers in breathing or the beauty of the bridge do not have exact numerical values. 2. but they have some features in common. violins. Some of these properties can be assigned numerical values while others cannot. Once this frequency is known. In almost all such cases. the amount of terror caused by the ride. cellos and double basses rely on vibrating strings to produce sound. The ﬁrst step is to ﬁnd. EXERCISES 1. but the problem needed to be dealt with.6: The Anzac Bridge With the Anzac Bridge. TRAGEDY AND BAD VIBRATIONS Figure 1. The properties may describe the system itself or its mode of operation. then measures can be taken to damp the vibrations. the atmospheric pressure and the frequency at which the cables vibrate are all numerical properties. in the three problems we have described. What factors do you think are signiﬁcant in determining the frequency of vibration of a string or cable stretched between two supports? Explain how these factors allow a stringed instrument to be tuned. Explain why a double bass sounds so different to a violin. It often happens that we may wish to calculate some numerical property of a system from a knowledge of other numerical properties of the system. On the other hand. calculating a numerical value will involve the concept of a function. In the mathematical analysis of problems. the vibrations caused annoyance rather than catastrophe. the natural frequency of the vibration of the cables supporting the bridge. For example.4 1. Musical instruments such as guitars. In each of them there is a complex system which can be described in terms of a collection of properties. 1. We will become quite precise about this concept in the next chapter.8 TERROR. but for the moment we shall consider a few special cases.5 DISCUSSION The three problems we have outlined are very different. we will only consider the numerical properties of objects or systems. ¥ ¦ . the time taken for the car to fall from to . at least approximately.
we have a similar situation. If it is displaced from its equilibrium position. we need a rule of calculation which relates two variables. two different values of may correspond to the same value of . but this may depend on the way in which the problem is posed. we can use a device known as an altimeter to measure height above the earth’s surface by observing the air pressure. in the 3 ¤ 2 £ .7: A vibrating cable In a given problem. We let be the height above the baseline. to put it another way. To ﬁnd the frequency of the vibration we need a rule which relates the displacement of the center of the cable to the time . In the second problem. so that is the dependent variable and is the independent variable. 1. while depends on the choice of .6 RULES OF CALCULATION Each of the above problems suggests that to get the required numerical information. For example. Here we are free to choose (as long as it is between and the height of the atmosphere). Let us try to distil the essential features of rules of calculation which can be deduced from the three examples we have presented: The independent variable may be restricted to a certain range of numbers. or the velocity at a given height . or the velocity at any time .7. In this case. the dependent variable is and we want a computational rule which enables us to calculate if we are given . which is taken to be the level of in Figure 1.2. since we are free to choose its value independently. since its value depends on our choice of . We may wish to know the height of the car at any time . We often call the independent variable. there is often a natural way of choosing which variable is to be the independent one and which is to be the dependent one. the car may be at the same height at different times. in the case of the Anzac Bridge. For instance. Finally. in the case of air pressure. There are many ways of arriving at such rules—we may use our knowledge of physical processes to deduce a mathematical rule of calculation or we may simply observe events and come to trial and error deductions about the nature of the rule. while height is the dependent variable. The quantities and are often referred to as variables . In the ﬁgure. the car cannot be in two different places at the same time. Or. while is called the dependent variable. Suppose we consider the variation of height with respect to time. we can represent one of the cables schematically as shown in Figure 1. the cable is ﬁxed under tension between two points and . it will vibrate. Here the independent variable is . This height changes with time and for each value of . ( ¤ % ) ) £ ( ) 2 1 ( 0 ( ( ) 2 ) ) ( ( 1 0 ) ) ( ¥ ) ( ) 1 ) ( 2 ( ( ) ( Figure 1. It is the value of which is given in advance and the value of which is then calculated. On the other hand. there is a unique value of . This is because at a given time the car can only be at one particular height. pressure would be the independent variable. We wish to ﬁnd a rule which enables us to ﬁnd the pressure at a given height . there are a number of possibilities. Notice also that the two quantities and play different roles.RULES OF CALCULATION 9 In the Tower of Terror.
9 4 112. The pressure on the hull of a submarine at a depth is .1 . 4. List three ways you could measure a variable such as terror and invent a function that uses terror as an independent variable. is the 2 83. TRAGEDY AND BAD VIBRATIONS case of the Tower of Terror we might not be interested in considering values of time before the motion of the car begins or after it ends. one of the most important ideas in mathematics. 0 # " 1 54. There has to be a procedure which enables us to calculate a value of the dependent variable for each allowable value of the independent variable. 2.4 3 102. ! In the next chapter. Year Population(Millions) 1950 10 1960 12 1970 14 1980 16 1990 18 Decide on suitable independent and dependent variables for this problem.6 1. However. we get one and only one value of the dependent variable. Hypothetical data values for the Tower of Terror are given in the table below. EXERCISES 1. rather than .5 6 102.6 5 112. Hypothetical census data for the population of a country region is given in the table below. In the text we stated that the amount of terror experienced by the passengers in the Tower of Terror could not be assigned an exact numerical value. What is a reasonable domain for ? 3.0 Suggest reasons why we have to take as the independent variable. This range of allowed values of the independent variable will be called the domain of the rule.4 9 14. Is there only one possible choice? Plot a graph of the data and use the graph to predict the population in 2000. Explain how we can regard either of the variables and as the independent variable and the other as the dependent variable. some people may be more terriﬁed than others.10 TERROR. Can you be conﬁdent that your prediction is correct? Explain. In the case of atmospheric pressure. we shall formalise these ideas into the concept of a function. nor should it extend to regions where there is no longer any atmosphere.6 7 82. where height at time .1 15.9 8 53. Suggest a reasonable domain when is the independent variable. the height must not be less than zero. ! To each value of the independent variable in the domain. so there is clearly something to measure even if this can’t be done exactly.
CHAPTER 2 FUNCTIONS In this chapter we will give a detailed discussion of the concept of a function. multiplying this by EXAMPLE 2. The number on which such a rule operates is called the input number . However. Let be the rule which instructs us to square the input number and then multiply the result by In symbols we write . As we have indicated. In symbols. EXAMPLE 2. We can illustrate this idea schematically in the diagram below. there are two possible values for the output number . The machine then produces the output according to the rule given by . E 7 C C B + ' 7 + + ' ' + ' 7 < = 7 7 + $ 7 $ $ + $ 5 2 2 . Let us denote the input number and the corresponding output numbers by .we require rules of calculation which operate on numbers to produce other numbers. the rule by letters such as Thus a rule operates on the input number to produce the output number . namely and . if the input number is . which we brieﬂy introduced in Chapter 1. then its square is and gives the output number as . 2. Thus. a function is a rule or calculating procedure for determining numerical values. A concrete example of such a machine is the common calculator.1 . by . is given by . the nature of the real world may impose restrictions on the type of rule allowed.2 Let be the rule which instructs us to ﬁnd a number whose square is the input number. while the number produced by applying the rule is called the output number .1 RULES OF CALCULATION In the problems we shall consider. % + ' % + $ + $ 3 1 1 % ' % $ $ Here we think of a function as a machine into which we enter the input number . Thus if the input number is .
Its natural domain is the set of all be the same rule. In the three problems that we considered in Chapter 1. while in Example 2. In these circumstances. It is often quite difﬁcult to determine the range. then is negative and in the context of this problem.1. if cannot occur. however. we take the value of the independent variable as the input number for the rule in order to generate the value of the dependent variable or output number. while is the output number or dependent variable. In the case of Example 2. Notice that there is nothing signiﬁcant about using E 2 C 2 2 E 2 C 2 2 2 £ 2 £ E 2 C £ i g 2 ` U " E 2 C 6 © g 6 f % g © w 6 .1 Functions Let be a set of numbers. It is sometimes useful to consider a rule in which the set of allowable values of the input number is smaller than the natural domain. we remarked that we needed rules of calculation to compute values of the dependent variables.2 negative input numbers will not produce an output number. or if . but restricted to the set of positive real numbers: in symbols . each output number is produced by two input numbers. is produced by and . EXAMPLE 2. Let 2 7 2 U E ` 2 C " © r % ) ) ) ( r ) ) % ( ) ( r w ) ) % u ) ) r f ) r g ( of Example 2. any number can be used as the input number. DEFINITION 2. reaches a maximum height before starting to fall and ﬁnally reaches the ground again at time . called the input number or independent variable. we are now able to give a precise deﬁnition of what we mean by a function. they nevertheless raise points which need clariﬁcation.1 given by . The new rule is called a restriction of the original rule and such restrictions may have properties not possessed by the original rule.12 FUNCTIONS As simple as these examples appear. For . The physical interpretation of the problem is that the particle begins rising at time . The second point about the above examples is the fact that there may or may not be restrictions on the values of the input numbers. so it seems that not all rules of calculation will be appropriate in practical problem solving. Every rule has a natural domain and to be a solution to a practical problem. number. Here is the input number or independent variable. We call the another number from and output number for or the function value at . each output number is produced by just one input is . The natural domain of the independent variable is the set of all real numbers. it is sensible to restrict to the values EXAMPLE 2.4 Consider the rule real numbers. You should think carefully about the meanings of the various terms in the deﬁnition of a function given above: is a number. is a rule for calculating is the number we get when we apply this rule to . For example. a rule must have the property that every physically reasonable value of the independent variable is in the natural domain. These problems always had a unique value of the dependent variable (output number) for each value of the independent variable (input number). The following examples illustrate some particular cases. The only input number for the output number With these considerations. a negative height However. The natural domain of a rule is the largest set of numbers which produce an output number. For . This is certainly not the case for Example 2. In these problems. enables us to calculate a single output number The set of all possible output numbers from a function is called the range of the function.2 above.3 Suppose that a particle moves so that its height above the earth’s surface at time is given by . A function with domain is a rule or computational procedure which for each input number in the set . .
We simply show the correspondence between the input number and the output number. In practice.1 deﬁned by . EXAMPLE 2.RULES OF CALCULATION 13 the letter to denote a function or the letter for the independent variable. ) ) for all . The independent variable is a number such as or . We can compute function is the set of all numbers for which either set of numbers. We read this as goes to (or Example 2. or any other letter. Find . it doesn’t matter what letter we use to refer to the function. We can even specify a function without using such letters at all. A function is some rule of calculation and as long as we understand what the rule is. EXAMPLE 2.7 Let Solution 2 g E 2 C . l u ) u % % k % where can be any real number. so that the natural domain of the or . Thus the function of may be written as . Various Greek letters such as maps to) or are also used. and . 7 by the rule . and it is irrelevant whether we denote it by . Its range is the same i You should convince yourself that and deﬁne the same function.5 We deﬁne a function 6 2 a ' g " U g ( 2 U 7 2 2 2 U E 2 C 2 6 ) 2 U ¢ 7 by the rule h ) h u d E E ) 2 g " " g 2 f " 2 u g C ) 7 d E ' d E ) 2 " l d 2 ) 2 ) g 2 C d g f ' ' E C E f d 2 E 2 E d d ) f g ' ) C C C C 2 g 2 2 g g g g g g g ! E E 2 E E E E ) C f ' 2 2 2 C C C g We also deﬁne a function In order to completely specify a function. This is called the natural domain of the function. it is necessary to give both the domain and the rule for calculating function values from numbers in the domain. Common letters to denote functions are and .6 We deﬁne a function by the rule 7 " " d d C d f C C E ' C " d 2 d EXAMPLE 2. In this case it is assumed that the domain is the set of all numbers which produce an output number when the rule of calculation is applied. it is common to give only the rule of calculation without specifying the domain.
Explain the meaning of the ( 13. 4. U g d f 2 2 2. g ' f r t E 2 " d " C d " E 2 C ( r 7 U 7 " f r t r t % %  ) ` w f 2 2 if if r t " 5. " ' E 2 C ' d 2 E 2 C 8. Use the principle of conservation of . but . E " d 2 C E 2 ~ " C 7 E 2 f C In Exercises 8–12. where m/s is a constant and is the mass of the carriage. which states that is a constant. calculate and be the vertical ' g ' r t g d 2 a " " f § 2 E 2 C 7. 2 ( k 7 ( k Explain why . ¤ 14. 6. The population (in millions) of a city is given by a function elapsed since the city was founded in 1876. ¤ ¥ ¤ (a) If the velocity at is 162 km/hr. Functions . 9. known as the principle of conservation of energy. on the line in 2 " ' f f § 2 2 ' E 2 C ( 7 2 " ' d d § 2 2 ' E 2 C 7 2 ' ( E 2 C 16. Do you think that it is meaningful to consider physical quantities for which no method of measurement is known or given? " { ' d 2 E { 7 2 C 17. if if " d 2 E 2 C " " d d 2 v 2 E 2 C 10. We have . Express the distance between the origin and an arbitrary point terms of .14 FUNCTIONS EXERCISES 2.2 on page 3. statement © U g E © a % " C © ¦ (b) Suppose the vertical distance of from energy to calculate the vertical distance ¥ is 14 m. what is the velocity at in m/s? whose input number is the time . 11. { g  ) " % d f v " ' E E E ) { 2 C C C . ﬁnd the numerical value of the function at the given values of . let be the velocity of the carriage at time and let height above at time . In Figure 1. and are deﬁned by the rules E ) C 15.1 In Exercises 1–7. ' r t E 2 C r 7 ' r t ' f 2 E 2 C r 1. The total energy at time is deﬁned to be ) 7 ( ) d 0 0 ) ¤ g g w 2 2 a " " f ! d § 2 2 g U E 2 C 12. 3. There is a physical law.
We call the union of and . We write Finally. % B v is deﬁned for all positive integers 2. Q Q r % r The closed interval r + ! is the set of numbers is the set of numbers K r satisfying r We also need to consider socalled inﬁnite intervals . Challenge problem : The following function as McCarthy’s 91 function. g g g g g K Q r denotes the set of numbers  satisfying Q denotes the set of numbers satisfying . Let . in which case in this case. Then denotes the set of all numbers for which or . h to denote the set of all real numbers. we use the notation to denote the elements of which are not also in . the intersection of and is said to be empty. . while K K  % The open interval % r + ! ! satisfying Q  + Q Show that for all positive integers . We call the intersection of and . then . while . . (c) The surface area of a sphere is proportional to the square of its radius. also allows to be an element of both We use to denote the set of all numbers for and . Express the following statements in mathematical terms by identifying a function and its rule of calculation. . or is an element of . % M Q if if  . When does equal and when does it equal ? and known r ! ! . It may happen that and have no elements in common.2 INTERVALS ON THE REAL LINE The domain of a function is often an interval or set of intervals and it is useful to have a notation for describing intervals. This is read as is in . .INTERVALS ON THE REAL LINE 15 19. Let be an interval on the real line. . let and be any two intervals on the real line. but this is usually denoted by A little set notation is also useful. It and . (d) The gravitational force between two bodies is proportional to the product of their masses and and inversely proportional to the square of the distance between them. we write Next. denotes the set of all denotes the set of all B  B d + +  +  +  B   + % 18. (a) The number of motor vehicles in a city is proportional to the population. If is a number in . Note that the mathematical use of the word “or” is not exclusive. 20. l % l B + r r % l B + r M l % r r r % l B + r l % r The notation numbers + denotes the set of all numbers denotes the set of all numbers The notation numbers K ! ! We can also use the special symbol . (b) The kinetic energy of a particle is proportional to the square of its velocity.
ﬁnd a function for which is the area whenever the input number is the circumference. 2. ﬁnd the domain and range of the function.4. both the domain and range are . Express your answer in the interval notation introduced in this section. E l 7 % C % w 2 . 2 g f " E 2 C " f 2 E l 7 % C E % 7 l f C is . that is. E l 7 l f C g 2 E ` 7 g C E U 7 " f C g 2 13. and . E " d 2 C 2 E ' f 2 C E 2 C 4. E g 7 " C g 2 E l 7 " C g 2 l 7 % " ! 7 7 " r C C g g 2 9. 18.9 and Let E U 7 " . E l 7 % E l 7 l f C 2 2 U U E E 2 2 C C 6 © EXAMPLE 2. 3. 16.1. Then . 21. . h g 2 w 2 ' u " 2 u " g 15. 10. E 14. 17. the domain is . 6. 11. Express the area of an equilateral triangle as a function of the length of one of its sides. " E 2 C  2  E 2 C 1.16 FUNCTIONS EXAMPLE 2. f 2 r w 2 and ' g  2  U u  ' d 2  19. What is the output number the domain of the function? 8. Rewrite the expressions in Exercises 9–14 as inequalities. .8 For the function For the function of Example 2. 2 ¥ E ¥ C Rewrite the following expressions in interval notation. while the range is of Example 2. 12.10 The natural domain of the function " 2 E 2 C EXERCISES 2. 7. ' 7 " E U 7 ' C E r 7 " E r 7 ' C EXAMPLE 2. Express the area of a circle as a function of its circumference. % 2 7 ' d 2 E 2 C ( " u 2 % 7 2 E 2 C 5. u 2 g 20.2 In Exercises 1–6.
Computers plot graphs in a similar way to the above example—they calculate many function values and then join neighbouring points with straight lines. but we shall not investigate these methods in any detail. but one method is simply to plot points until we can get an idea of the nature of the graph and then join these points. then its graph is the set of all points of the form .1: The graph of E 2 C . In many cases there are shorter and more elegant ways to plot the graph. With the advent of computer software such as Mathematica and Maple. It is often easier to interpret information if it is presented visually. It is often difﬁcult to sketch graphs. which we have plotted using the software package Mathematica. EXAMPLE 2. 15 10 5 4 2 2 4 Why do we draw graphs? One of the main reasons is as an aid to understanding. The essential things that Mathematica needs to know are the function to be plotted. the need to plot graphs by hand is not as great as it used to be. 2.GRAPHS OF FUNCTIONS 17 2.3 GRAPHS OF FUNCTIONS If is a function with domain . This is shown schematically in the diagram below. the straight line segments joining them are very short.1 Plotting graphs with Mathematica The instruction for plotting graphs with Mathematica is Plot. where is any number in (written ). We tabulate values of for various values f 2 g E f 2 C 2 C . The procedure for any other example is no different—it may just take longer to compute the function values.3. Because the plotted points are so close together. as we have done in Example 2.11 Let the function be deﬁned by the rule of : ' f U f 2 2 2 E E 2 C 7 2 C ¦ g ¦ 2 ¦ 0 " 1 1 2 4 3 9 4 16 16 E 9 4 1 0 If these points are now plotted on a diagram the general trend is immediately evident and we can join the points with a smooth curve. and the overall impression we get from looking at the graph is that a curve has been drawn. rather than as a formula or in tabular form.11 below. the independent variable and the range of values for the i 2 E 2 C Figure 2.
5 10 7. McLelland.15. University of Technology. There are numerous optional extras such as colour. current edition of this book has not been printed in colour.18 FUNCTIONS independent variable. 1 EXAMPLE 2. Notice that Mathematica may sometimes choose to override our instructions. However. 1996.12 To plot a graph of the function given by Y . axis labels. 2 The 1 For + Y E C {  + 2 4 .4}.PlotStyle>{RGBColor[1. Introduction to Mathematica by G.0.0]}. In this case it has not put a tick for . Ticks>{{4. J. Sydney. we use the instruction Plot[xˆ2. but we will assume that you are familiar with the basics of Mathematica. while AspectRatio alters the ratio of the scale on the two axes.5 5 2.4}.0. green and blue.{x. Use the Mathematica instruction given in the text to plot the graph of . frames and so on.{5.AspectRatio>1] Here the Ticks option selects the numbers that we wish to see on the axes.2. example.1. the curve is rather squashed and there is no colour.1 (without the dot points).4. or have access to supplementary material. we can add a few more options. 2 EXERCISES 2.4. To improve the appearance. In this book we will explain how to use Mathematica to perform certain tasks.2: A basic Mathematica plot This graph does not look as nice as the one in Example 2.20}}. The numbers in RGBColor gives the ratios of red.4}] This will produce the following graph: 15 12.5 4 2 Figure 2. we used the instruction Plot[xˆ2.3 1.2.{x. The vertical axis is too crowded.10. To plot the graph in Example 2. you can get colour graphs on a computer by following the given commands. In this case the graph is 100% red. Experiment with different colours and aspect ratios to see how the appearance of the graph changes.
14.x] f3=Fit[a.9 .x] will give a formula for the best quadratic and best cubic polynomial respectively ﬁtting the given points.4 .GRAPHS OF FUNCTIONS 19 2. we can use the Mathematica command Fit. Notice that these points lie on what appears to be a smooth curve. ﬁnd f2.x. Here is a table of values which could apply to the Tower of Terror.5 102. 0 1 2 3 4 5 6 7 8 9 15.83.6 . (b) Use the Mathematica command ListPlot in the form g0=ListPlot[a.15 . 1. The command f1=Fit[a. 9.53.x.5 s. This command ﬁts a curve of the user’s choice (straight line.x] will give a formula for the best straight line ﬁtting the given points. Use this formula to calculate the height at each of the times given in the table.4 14.54.1 83.x 2.PlotStyle>PointSize[0.5 . 1. the commands f2=Fit[a. (c) Rather than estimate the curve by joining the points by hand. the dotted plot is the graph of .112. 7.9 112. that is. 5.6 82.9 .6 .1 (a) Use the command a= 0. 2. Plotting a graph over a part of its domain can lead to deceptive conclusions about the nature of the graph in its entire domain. 4.1 to enter this table as a list in a Mathematica notebook. quadratic polynomial and so on) to a list of data points.02]] to plot these points on a graph. 3.5 seconds.x 2 . Does this agree with your answer in part (b)? ¡ ¢ ¡ ¢ ¢ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ . Print a copy of the graph. 8. What would you have concluded had you plotted only one of these graphs? C + + Y 3. obtain the following plots for and .82. join the points by hand with a smooth curve and use this curve to estimate the height after 2. 1. Fit a quadratic polynomial to the points in the above table.x 3 .1 . How good is the agreement? What conclusions can you draw about the data? Determine the height after 2.112.4 .9 53. By selecting a suitable interval.102. There are numerous options and you should consult the Mathematica help ﬁles for a full explanation of this command.0 54.102.6 112. 6. Similarly. In each case. 1.4 102.x .
0. use the formula to calculate the pressure at the heights given in the table and compare the results with the tabulated values. (c) Fit a straight line. 6. Studying and doing many examples exposes you to many different aspects of a concept and should help your learning. Notice that these points lie on what appears to be a smooth curve. This implies that when we use a function to calculate the value of the dependent Y £ . 5. there is one matter which needs to be clariﬁed. Plot a graph of this area as a function of and decide from your graph at what value of the area of the rectangle is a maximum. is known or welldeﬁned ? In this book. In each case. Very often. A rectangle of height is inscribed in a circle of radius . we emphasise the use of functions to compute numerical data relating to problems in the real world. Before giving these examples. and then display the graph g2 and the plotted points g0 simultaneously by using the instruction Show[g0. plot the graphs of the functions obtained in part (c). Plot a graph of the distance of the airplane from Melbourne against time from the moment of takeoff to the moment of landing.20 FUNCTIONS (d) You can plot the formula f2 with the instruction g2=Plot[f2. x. a quadratic polynomial and a cubic polynomial to the set of points in the table above. join the points with a smooth curve and use this curve to estimate the pressure at a height of 4500 metres. Here is a table giving the atmospheric pressure at different heights: 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10 000 1013 899 795 701 616 540 472 411 356 307 264 ¡ (a) Enter this table as a list in a Mathematica notebook (b) Plot these points on a graph.4 EXAMPLES OF FUNCTIONS In this section. Print a copy of the graph.g2] 4. A airplane ﬂying from Tullamarine Airport in Melbourne to KingsfordSmith in Sydney has to circle KingsfordSmith several times before landing. Find an expression for the area of the rectangle. and that is the one posed by the following question: What do we mean when we say that a given function . say.10 ] Do this. (d) On the same set of axes. we consider various examples of functions. Which curve gives the best approximation to the table values? Compute the pressure at a height of 4500 m and compare it with your result in part (b). 2. mathematical concepts possess subtleties which are not immediately apparent. Looking at different examples is a useful learning method in mathematics and one which you should try to cultivate.
Now let us turn to some examples. A diagram such as Figure 2. but it does not tell us certain critical properties.3. enables us to compute to any For example.) 2. Accordingly. In describing any function. For example. ¤ EXAMPLE 2. These consist of sums of the form where r E % r ` a E ! 2 C ` g 6 ` " C 6 6 2 U E 2 C 6 6 2 E 2 C are constants for a given . the answers we get should agree with the experimental data. Computing an output number such as to any degree of accuracy can be done. There must be a rule which produces one (and only one) output number for each input number in the domain. 7 r d 2 ` r d " d ¦ g ¦ ¦ d U d ` a ¥ " ¤ 2 f a " ¥ g ` ¤ g % r f d ¤ 2 ¤ r E 2 C r 7 7 r 7 . by using a graphics package such as Mathematica. The rule must be in a form which enables us to compute the output number to any speciﬁed degree of accuracy. if we are able to in decimal form to an arbitrary degree of accuracy.EXAMPLES OF FUNCTIONS 21 variable in a practical problem. obtained using Mathematica.1 given by number of decimal places that we choose. more technically. multiplication and division. the only values which are measured in the real world are decimal numbers to a certain accuracy.3 may give a good idea of the appearance of the graph.1601265326969. one answer to the question posed above is: We say that the function is known or. for any in the domain compute of . The domain of the function must be known. are there any turning points of the graph besides the 3 The i h h exact answer is 12968963. (But see the remark on page 13 about natural domains. subtraction. more conveniently. One simple way of deﬁning functions is by arithmetic operations—addition. If the rule does not allow us to do this.13 A function deﬁned by the rule is a polynomial function. As far as such data is concerned. although it may take some time. The graph of the function is shown in Figure 2. for example ` d " g d " g § U f § " g E " g C ` d 2 d 2 U f § 2 E 2 C The (natural) domain of the function is and it can be shown that the range is also . the rule that deﬁnes a function should be able to produce output numbers in decimal form with arbitrary accuracy. 3 Thus is a welldeﬁned function. Since more sophisticated methods of measurement may increase accuracy. we must insist on three things: 1. is welldeﬁned. We can compute its value for any input number. and the simplest functions of this type are polynomials . We can plot the graph either by plotting points by hand and joining them or. the rule of Example 2. then we cannot really say that the function is known. 3.
3: The graph of P ' g " ¨ r r f " f g f .14 We deﬁne a function 1. as the following example shows. the domain of is . g " u 2 % In symbols we can write 7 % u 2 h As we have already mentioned. . In Figure 2. then: % u 2 The output number is 1 if d 2 g 3 . the domain of a function is an inﬁnite interval. In the above example. EXAMPLE 2.22 FUNCTIONS ones shown? In many cases of interest. that may occur in the region we have not plotted. The rule for computing function values obvi 7 7 " g u u g 2 2 " % 2 E " " f d 2 C 2 " g ' ' g «© « « « « « ¬ 2 The output number is 3 if 3 . we have speciﬁed the rule for calculating function values by giving a formula which applies at all points of the domain. the rule for calculating function values may be given by a different formula at different points of the domain. in which case we can only plot part of its graph. However. 2. The domain of 2 by the following rule: h is . such as turning points. If is the input number. u 2 " E " f E 2 C 2 C ' The output number is 3 " The output number is 3 if if .3 all the turning points of the graph have been plotted. We cannot then be sure about features. but we would need to do some further mathematical analysis to show this. if if if if ` d 2 d 2 U f § 2 E 2 C Figure 2.
Notice . Here denotes the set consisting of the two and 1. so these numbers are excluded from the natural domain.4. such as domain is the set of all real numbers other than or .15 Allowing the operation of division in the rule produces more complicated examples. Here we are using the common convention that this circle the small circle centred at the point indicates that the point at its centre—in this case —is excluded from the graph. the natural and . % % + % ¨ + % + EXAMPLE 2. elements Numerical evaluations of the function values can be carried out by hand. The result is shown in Figure 2.5. This is written is various ways. while denotes the set with the two points and 1 removed. and are not part of the graph.EXAMPLES OF FUNCTIONS 23 ously produces numerical values of any accuracy: B H = + H = + a The graph of the function is shown in Figure 2. Since we can use any other number in the above expression for . Let us deﬁne a function by C C B B B d + Y Y Notice that the denominator is zero for and .4: The graph of 2 7 a B H H C H H d C = C H C C C B H H C + B a + a C B C B . but they become tedious: + Y C B ¡ C % B C H B h B H C C B C C B l % H + + Y ¡ % % C C B B + h C B % l B + C ¡ B % C B h To sketch the graph. it is easier to let Mathematica do all the work of evaluating function values and plotting the curve. E C B  + C = H B { Figure 2. We have also centered a solid disk at the point to indicate that this point is included in the graph. From the graph we see that the range is . They are called vertical asymptotes The dotted lines and we had to give Mathematica special instructions in order to include them on the diagram.
We give an example of such a function. The why of this example will be clariﬁed in Chapter 6. We are asking you to go through the same sort of unthinking process that a computer uses in calculations—just follow the rules. Unfortunately (or perhaps fortunately. ¡ 2. but only how to calculate its value for any input number. we deﬁne g g f f 2 2 f d 2 2 E 2 C Figure 2. but in most applications the functions we obtain as solutions have to be deﬁned in other ways.5: A Mathematica plot of the graph of . at ﬁrst sight. If you have not seen this rule before. EXAMPLE 2. describing phenomena in the real world is usually a difﬁcult procedure and if we are to make any progress we need some nontrivial mathematics. Let the input number be any positive real number. if it is properly speciﬁed we should be able to follow it and produce an output number. In fact the rule is very useful in applications and a large part of our later work will be to ﬁnd out why such rules are needed in practical problems and how we can obtain them. At this point we are not going to explain what the function is used for. whose rule of evaluation may seem strange. 3. Calculate each of the numbers 7 7 ¯ ° U 2 7 ¯ § ' 2 7 ¯ g 2 7 2 7 " ¡ ® 7 ¯ ¡ E " ¥ f ¤ 2 C ¥ ¤ E " f C d ¦ ¦ ¦ d ¯ § ' 2 f ¯ g 2 d 2 f " 2 4.24 FUNCTIONS 30 20 10 4 3 2 1 10 20 1 2 3 Functions deﬁned by algebraic formulas arise in some simple applications. depending on your point of view). " § g § ' § ¦ ¦ ¦ § ¤ E g f ± C § E " f ± C § ± ¯ ± " ± where for any . Select a desired accuracy of calculation . artiﬁcially contrived and of no practical use. we want the calculation to give an output number with an accuracy of . Suppose this number is ¡ . unusual. The purpose of this example is to show that no matter how complex or unusual a rule may seem.16 We are going to deﬁne a function by a rule which is. you probably will have no idea of what the function is used for or where the rule comes from. Here is the process: 2 1. It would be very easy to solve problems if the only functions we ever needed were nice simple ones. Continue until . Then the output number for E 2 C ¤ can be calculated with an error of less than from the expression ¯ 2 one of the numbers is less than . that is.
Then we ﬁnd .6. ﬁnding the range of a function can be difﬁcult. that is. we get the graph of i g g % E r " C E 2 C U g g % E r " C g g g % The message of this example is that if a function is properly deﬁned. plot them.EXAMPLES OF FUNCTIONS 25 where we have worked to 4 decimal places. we get U ' % % f ! r " % % d ' ' ` % % f % " g % d r g ` r % f r g " " d r " f " E r " C We conclude that . a number we can plot on a graph. In the case of rational functions with quadratic terms. to the stated number of decimal places. equal to the number on the left hand side. . We will see in Chapter 6 that a function of this type is needed to produce the solution to the air pressure problem we considered in Section 1. then we can compute its values. carefully specify the rules that enable the computer to calculate function values. The notation is used to indicate that the number on the right hand side is.3. and shown in Figure 2. let us turn to the problem of ﬁnding the range of a function. even if we have no idea where the function comes from or what it is used for. If we were to repeat this calculation for many input numbers. When we program a computer we have to engage in a similar process. Following our rule. we can sometimes use our knowledge of quadratic equations to ﬁnd the range. . interpolate a smooth curve. E 2 C { Figure 2. To 2 decimal places we have .6: The graph of 7 ` % % " % g % % % P ¯ ¯ µ ° ! U 2 2 U r U g ' % ` r % ' % % ¯ ¯ ´ § a ' 2 2 g ! r r " " g % " % % " % % " ¯ ¯ ¡ g ` 2 2 r r " ' g ' ' g g ¨ " ` g g r % % % 2 As an illustration. A rational function is one that can be expressed as the quotient of two polynomials. As we have mentioned earlier. take " 2 and ¡ d E 2 C ¤ u E 2 C u ¡ % f E ¯ 2 ³ C r 2 ¤ that is. Finally.
) ¹ { 2 { ¹ f " · E { C " · f 2 For every value of . Let % % % be the function deﬁned in Example 2. However. we found . This is a quadratic equation in and we need to ﬁnd the values of for which this equation can be solved to give real values of . ' h and this is welldeﬁned for all . Hence verify that the graph of is consistent with on a graph. does this mean that you don’t really understand ? Comment. In this example. Consider the equation ( ) { and We say that and are functions implicitly deﬁned by the equation ( ). 4. g ® k 2 2 It is easy to see that the natural domain is the set of all real numbers range. Can you give a computational procedure for calculating square roots without using tables or builtin calculator functions? If your answer is “no”. but this does not give rise to new functions: the function is the same function as . the functions and are not the only functions implicitly deﬁned by equation ( ).6. Evaluate . ' f E C 2 " f ) · E ) C 3. " d 2 8. 6. 2 f § 2 E " 2 d 2 E 2 C 5.4 In Exercises 1–6. together with the value of the one given in Figure 2.17 Find the range and natural domain of the function 2 2 2 2 C f E g g r C d ¹ E { " E 2 2 U C { C C 2 2 ¶ C and rearrange the equation to get . The discriminant is 2 2 { ` " % d " f 2 f 2 { { EXERCISES 2. Thus. namely .16. each of the above equations deﬁnes a function of according to f " { 2 { f " · 2 2 f " f and ® ¹ ' g g % E r " C " { E E E " C E r % C E % C " 7. Plot these values . (The matter of computing square roots what is meant by an expression such as will be taken up in Chapter 10. and with an error of less than 0. We can also solve the equation ( ) for in terms of . we can ﬁnd two possible values of . we put U f 2 2 U f 2 f 2 E 2 C for which . Can you think of some others? 9.001. Hence the range of { is . ﬁnd the natural domain and range of the function. To ﬁnd the . . 1. U f 2 E 2 C 2.26 FUNCTIONS EXAMPLE 2.
The function shown in Figure 2.1: Displacement of a moving body Á H 7 C À C H 7 Á H 7 ÂÃÄ C À C H 7 .1 SMOOTH FUNCTIONS What does it mean for a function to be smooth? In order to reach an answer to this question. we gave some examples of functions. 3. but in fact functions of this type have many uses. What can we say in general terms about the graph of ? Consider the two graphs in Figure 3. We do not expect to see bodies jump from one place to another in no time or to instantaneously change their velocity. the processes which occur in the real world happen smoothly.5 would rarely. occur in a practical problem.4 may also seem to be in this category. although we shall not consider them in any detail in this book. if ever. There are of course processes which do occur suddenly. However. In general. hitting an object with a hammer or perhaps falling off a cliff. The functions which occur in the problems in this book will all be smooth. We would not expect the air pressure at a particular height to change suddenly from one value to a completely different one. Suppose a body moves along a straight line with its displacement from a ﬁxed point at time being given by for some function . such as switching on a light. » º Y + Y º Y Figure 3. but these are really smooth processes which occur very rapidly on our time scale. let us consider some particular cases. In Chapter 2.CHAPTER 3 CONTINUITY AND SMOOTHNESS In this book we are mainly interested in functions which provide the solutions to mathematical models of real situations. not all of these will occur in practical problems.1. A function such as that of Figure 2.
The essential idea of continuity that emerges from these examples is that the graph of a smooth function should have no gaps or jumps. There is another aspect to the idea of smoothness. The left hand graph of Figure 3. because there is no point on the –axis is that at the time corresponding to the point seconds. First there is the case where a body moves with constant and a speed in a straight line. which we illustrate by considering two different cases for the velocity of a moving body. This is what we have to avoid if we are to rule out instantaneous changes in velocity. undergoes an instantaneous change of velocity . Loosely speaking. getting closer to a distance of 3 metres from as gets closer to 3 seconds.2: Displacement of a moving body As we have mentioned earlier.2. When .28 CONTINUITY AND SMOOTHNESS Let us look at the ﬁgure on the left. the graph represents a body which moves away from . Up to the time the graph will be a straight line of slope . the graph changes instantaneously to a straight line of slope 1. then plot of against gives a straight line with slope . the body is 3 units from . » º » » Q » » » Figure 3. so that a smooth function is one which is both continuous and differentiable. The graph on the right is more reasonable. the body has no position. from 0. a function is differentiable if its graph has no corners. at time . Now imagine a body which. while the one on the left is discontinuous. The physical interpretation of such a graph seconds. The noticeable feature of the right hand graph is the corner at .3 shows the case . a graph constructed C º º C % C » + C C C C º C % C C + º Á H 7 C À C H 7 . continuous or neither except to look at its graph. Here the distance from increases without any sudden changes in position. We have at present no way of determining whether or not a function is differentiable. We shall later show that all differentiable functions are continuous. Immediately after 3 seconds have elapsed the body instantaneously moves to a position only 2 metres from . The obvious difference between the two graphs is the jump in the ﬁrst one. At . Such behaviour of moving bodies does not occur in the real world and we want to investigate the restrictions needed to rule it out. For seconds. suppose a body moves along a straight line so that its displacement from a ﬁxed point at time is given by the graph in Figure 3. such behaviour is not in accord with our everyday experience. Next. the small circle centred at the point indicates that the point at its centre—in this case —is excluded from the graph. We would like to exclude such displacement–time graphs from our considerations. Such difference in behaviour is described by the notion of continuity : the graph on the right hand side is continuous. This behaviour is another example of lack of continuity and is quite unrealistic. The right hand graph of Figure 3. If denotes the distance from a ﬁxed point at time .5 m/sec to 1 m/sec. We often refer to differentiable functions as smooth functions. However.3 shows this case. Clearly.
It turns out that a careful look at the concept of smoothness will point the way towards methods for ﬁnding such functions.4: Two graphs through the same plot points As is usual in mathematics. Figure 3.SMOOTH FUNCTIONS 29 Figure 3. Historically. we will have to give precise and unambiguous deﬁnitions of the concepts discussed above. Our main aim is to ﬁnd the function which relates the variables in a practical problem. To know whether or not a function is smooth. There are also cases where smooth functions change in very short time periods. ¨ Figure 3.3: Smooth and nonsmooth displacements simply by plotting points is inconclusive. The income tax scale is one example. but one function is smooth while the other is not.4 shows an example where the graphs of two functions pass through the same plot points. There is a second reason for needing these deﬁnitions. It is often simpler to approximate such smooth functions by a nonsmooth function. we need methods which will give conclusive answers and to do this. because anything might happen between the plotted points. such deﬁnitions did not come out of the clear blue sky. This may occur when we model the turning of a switch on or off. We will not consider such applications in this book. we will need deﬁnitions to tie down the exact meaning of the concepts of continuity and differentiability. We should remark that in some applications we need to make use of functions which are not smooth. or 2 Å Á H 7 Å C H Å À C H 7 Á H B Å 7 H B C Å À C H 7 B Å .
we see that the jumps or gaps in the graph occur at speciﬁc points in the domain and we describe this by saying that continuity is primarily a point property. Sketch the graph of on the interval . There are several criteria—not necessarily independent of each other—which have to be satisﬁed if a function is to be continuous at some point of its domain.2 CONTINUITY Looking at the graphs of the functions in Figures 2.5: A graph with a gap  + Y 2 {  % ÆÇ K K < B Q if if C % ¨ v a a + a a h g . until some example arises which forces us to take more care with their deﬁnitions. This function is often used to model a signal generated by repeatedly ﬂicking a switch on and off. In our case. Do you think that the function is an accurate representation of the physical process? 3. We will motivate such criteria by considering various examples. 2. 3. A function may be continuous at some points and not continuous at others. The square wave function is deﬁned on the interval Q by % C < We extend the deﬁnition of to the entire real line by letting + a C d + a for all .30 CONTINUITY AND SMOOTHNESS inscribed on stone tablets. there is a need for careful deﬁnitions of continuity and differentiability to ensure that the functions we use to describe real world processes are in accordance with experience. Figure 3.1 1. they developed slowly over time in response to some perceived need. EXERCISES 3. Often concepts are taken for granted or are not properly understood.2.4. we begin with the idea of continuity at a point.5. In attempting to formulate a precise deﬁnition of continuity.1 and 3. Rather.
We can now state a preliminary deﬁnition of continuity at a point. + For our next example.6. . to 2 (the value of of its domain. Clearly. + Y Criterion 1 In order for a function to be continuous at a point must be in the domain of . to ensure that continuity.6: A jump discontinuity example. consider the graph of the function shown in Figure 3. while the values of for slightly less than 1 are close to 3 and the values of for slightly greater than 1 are close to 2. Clearly. that is.5. However.CONTINUITY 31 To begin with. the only way to avoid having such a gap is to ensure that there is indeed a is welldeﬁned. This gives us another criterion for continuity at a point. + Y + Y + Y Y + Y + Y + Y + C Y + Y + Y 2 7 C ¨ C B C B  + The graph of at the point is shown in Figure 3. For the moment we will continue at an intuitive level. In this % % C K K % C K Q Q if if if if a Y  d d + C Y «© « « « « « Y ¬  Y + a a . The word “close” is not very precise: how close is “close”? We can sharpen the idea a little by should be arbitrarily close to the value of provided is sufﬁciently saying that the values of close to . There would be no jump if values of were close ) for all close to 1. In this case. let us consider the function Figure 3. the function is not deﬁned at the point and this is manifested by the gap which appears in the graph above the point . so that we need a condition that describes this type of jump behaviour. the number must exist. given by . we would not want to consider this to be continuous . that is. This gives us our ﬁrst criterion for function value at . it is necessary that Criterion 2 In order for a function to be continuous at a point the values of are close to the value of as long as is close to . in our ﬁnal deﬁnition of continuity we will have to be more careful about the meanings of the various terms.
If the sequence is not convergent. we can compute the number as long as . since it makes use of the idea of sufﬁciently close. Some examples of sequences are given below. . there must be a rule for computing its terms. a sequence is usually thought of as a list of numbers rather than a function. then the sequence is said to be convergent and the number is called the limit. and . then it is said to be divergent . 2 are arbitrarily close to the value of 2 7 ` " 7 . The values of to . The most important property of a sequence is whether or not there is some number to which the terms get arbitrarily close as gets larger.1 (a) " 7 " f 7 " In order for a sequence to be welldeﬁned. that is. EXAMPLE 3. Then 7 7 E ' C 7 E g C 7 " E " C Despite the above deﬁnition. Notice that in each sequence there is one term for each positive integer and a rule for generating this term from the integer. a sequence is an inﬁnite list of numbers.2 A sequence of real numbers is a function whose domain is the set of positive integers. This leads to a precise deﬁnition of a sequence. The values taken by the function are called the terms of the sequence.1 Preliminary deﬁnition of continuity at a point The function is continuous at the point in its domain if: E 2 C 2 1. There are a number of common ways of indicating that the sequence converges to : 2 2 ¡ ¤ 2 i 7 ° 7 § 7 7 " so is a sequence and its terms are i § 7 Ê " U d " É 7 Ê " ' d " É 7 Ê " g d " É 7 Ê " " d " É (d) ° § 7 ° 7 § 7 7 " (c) 7 " f 7 " 7 " f 7 (b) 2 E 2 C 2.2 Deﬁne a function where ¡ ¡ ¤ E ~ " d " C ¡ ~ " ¡ ¤ E " f C by 7 ~ " E C is a positive integer. If this is the case. Loosely speaking. EXAMPLE 3. A sequence is often referred to by giving the general term in braces. One way to clarify this is to use the idea of a sequence. Thus the above four sequences are . DEFINITION 3.32 CONTINUITY AND SMOOTHNESS DEFINITION 3. The function is deﬁned at the point E 2 C The above deﬁnition of continuity deﬁnition lacks precision. is sufﬁciently close 7 U 7 " .
3 Continuous functions in its domain if the sequence The function is continuous at the point for every sequence in the domain that converges to . It is not obvious what happens from the terms given. The sequence (c) converges with limit .3 of Example 2. so that is deﬁned for all . this particular sequence is important. 2 2 ¡ ¤ 2 E 2 C converges to ¡ Notice carefully the words “for every” in this statement. it is enough to ﬁnd just one sequence converging to such that the sequence does not converge to . as tends to inﬁnity. Notice that we have implicitly assumed that Criterion 1.5 of Chapter 10 we will show that it is convergent with limit 2.CONTINUITY 33 In simple cases we can determine whether or not the sequence converges by looking at the list of terms. If we take any sequence in the domain of which and then consider the corresponding sequence of function values . we . since the terms get arbitrarily close to as .1. DEFINITION 3. The sequence (d) is more difﬁcult. we can improve our understanding of the idea of continuity of a function at a point in its domain. % 2 . this latter converges to sequence must converge to if is continuous at . In each case. However. 2. 2 ! " ¤ a 2 g Ó 2 Ê d " ¡ É ¤ l Ô 2 Ñ Ï Î Ð Í E ¤ 2 l C 2 2 ¤ 2 We may write 3 as .71828 .44. In Theorem 10. 2.25. We read this as approaches 2 ¡ ¤ 2 We may simply state the fact in words: 3 converges to . a number commonly denoted by the letter . If we compute the decimal approximations of the ﬁrst few terms we get 2. say. One of the simplest is terms are ¡ ¤ 2 7 7 ° g 7 § g 7 g 7 ' 2 U E 2 C 6 2 ¡ ¤ 2 E 2 C 2 ¡ E ¤ 2 C 2 . as well as those which oscillate from side to side. To investigate the continuity at Consider the function need to consider input sequences which converge to 2. the sequence (b) diverges because the terms oscillate. that is. we assume Ó Õ ¤ " With the idea of a sequence available. This is another and more precise way of exists. EXAMPLE 3.37.1. The ¡ ~ ¡ E ¤ 2 C g 2 " ¡ d E ¤ g 2 2 C E ¤ 2 C % ¡ 2 ¤ E 2 ¤ 2 C 2 Ñ ! Ï Î Ð g Í A more compact way of writing the statement above is ¤ 3 . to show that a function is not continuous at a point . There are methods for dealing with such sequences and while these are important for many practical purposes we shall not need them in this book. and of course each number in the sequence must be in the domain of the function. the sequence (a) diverges because the terms get larger without any bound. For the present we shall assume this result without proof. However. We must consider input sequences on both sides of the point .49. there is no number that the terms get close to as gets large. It is still not clear what will happen. which is required by stating Criterion 2. This property of convergence must hold for every sequence which approaches . 2. Thus in Example 3. 2.
Since In this last example we have made assumptions about the convergence of combinations of sequences. There is a theorem which deals with these matters. which is the function value at 2. Hence converges to . so the output sequence converges to . For example. we assumed converges to if converges to . The terms of this output sequence are . Let’s have another try without tying ourselves down to a particular input sequence such as . Then it is reasonable to suppose that converges to and hence that converges to . EXAMPLE 3. and we have only tried one sequence. It is often possible to use this theorem to prove results on continuity.1 Let be a sequence converging to 2 ± ¡ ¤ 2 ± ¡ E E i d ¤ g g C C g 2 6 E ¤ 2 C 6 g Ñ Ï § Î Ð Í ¡ ¤ E ¤ U 2 C 6 g 6 ¤ 2 g Ñ § Ï Î Ð Í ¤ U ¡ ¤ g 2 2 ¡ ¤ 2 U 6 ¡ ¤ g 2 ¡ ¤ 2 ¡ ~ " This theorem can be proved using a careful deﬁnition of convergence of sequences. We state it without proof. we note that " f " d It should be clear (and this can be proved) that 7 % " Ï Î Í " " " " d % d d k f ¤ { Ô " d " Ñ " % Ð d ¤ k { { ~ 2 ¡ ¤ { ~ ¤ 2 (e) converges to { 2 ¡ ¤ { ¤ 2 (d) converges to if { f 2 ¡ ¤ { f ¤ 2 (c) converges to { d 2 ¡ ¤ { d ¤ 2 (b) converges to and ± (a) converges to for any constant for any . THEOREM 3. but we shall not give the proof here. Then ¡ ¤ 2 . We have shown that there is a sequence with for which . { ¡ ¤ { and let be a sequence converging to 2 E § d g C ¡ U 7 ¤ 2 E d g 2 C U 7 2 E " d g C ¡ ¤ U ¡ 2 E ~ " g d which converges to g C U ¡ E ¤ 2 C 6 as tends to inﬁnity. ? Deﬁnitely not! The deﬁnition of continuity requires Can we conclude that is continuous at that the sequence of function values converges to the actual function value for any input sequence. can be any sequence converging to 2.4 In order to decide whether or not the sequence Õ ¡ ¤ r converges. Let be any sequence which converges to 2. The sequence of function values or output sequence is .34 CONTINUITY AND SMOOTHNESS The terms in brackets converge to 2. we conclude that is continuous at 2. as we will show below.
We have looked at two input sequences. Let ¡ ¤ + Y Õ d ¡ Ô ¤ + + Y We see that as . We have Ö v as We see that as . so that the function is deﬁned at 3. Let Õ + Y ] ¡ ¤ + Y Then is a sequence converging to 3 from above. Next. Ê d Ñ B Ð ¤ ¤ . Next. but we have not function values converge to proved this.5 Suppose we want to investigate the behaviour of the function given by the rule near the point . We may be interested in investigating properties of the function such as continuity or differentiability at the point . it is often useful to take a sequence converging to and see what happens to the sequence . To this end. we choose an input sequence that converges to 3 from above and see what happens to the corresponding sequence of function values.CONTINUITY 35 so that E Y l ] B É Ï Î Í Ï ] ¡ ¡ Y ¡ ¡ ¤ ¤ ¤ ¤ d + + l Y l Y Ñ ] ] Î Ð Í + Y + Y ] ¡ ¤ + Y ¡ ¤ EXAMPLE 3. First. note that . which converge to 3. we choose an input sequence that converges to 3 from below and see what happens to the corresponding sequence of function values. Õ l Ê d ] < B as Õ Ö Ê d < d B d B É É Ô v Ô Ô ] ¡ ¤ Then is a sequence converging to 3 from above and we have ¡ ¤ r Hence the sequence converges to 1. In both cases the corresponding and it seems plausible that is continuous at .
so by part (a) again. Hence. converges to .36 CONTINUITY AND SMOOTHNESS In the next example. which converges to 1. We do this by taking the input . Show that is continuous at . it follows that is continuous at . Note. we show how Theorem 3. by Theorem 3. but they are not always the best tools to use in practice. however. so the output sequence does not converge to the function value . E ¡ ~ ¡ B ~ ¡ d ~ + B + Y ¡ ~ B ¡ ~ d % % C K K C Q Q if if if if Y ¡ B ¤ « « « « « ©« ¬  + Y  + Y ] ¡  ¤ + Y so that ous at 3. We can prove theorems using the deﬁnitions and it is these theorems which provide the practical tools. which converges to 0. and choose any sequence converging to 3. converges to by part (d). Since this is true for any input sequence converging part (b).8 Deﬁne a function Y H ¡ ¤ H ¡ ¤ H d H d + ¡ Y ¤ ¡ H ¤ ¡ d ¤ ¤ H d + Y by % K We shall show that this function is not continuous at . These examples illustrate the roles played by deﬁnitions and theorems.8. Deﬁnitions provide us with the concepts and state precisely what we mean by the concepts. however.1 may be used to prove continuity. However. To show that a function is continuous at a point . however. the output sequence does converge to the function value. The output sequence is also . Then using Theorem 3. to . for any sequence converging to 3. we need to show that the output sequence converges to for every sequence which converges to .6 . Let be a sequence converging to . Let Solution.1. conclude that it is not continuous on the basis of just one input sequence. This happened with Example 3. Solution.7 Show that is continuous at any point . that if we had taken the input sequence and the function is not continuous at . the sequence function value at is . Notice. EXAMPLE 3. EXAMPLE 3. by converges to . Then. We can. This illustrates the fact that we cannot conclude that a function is continuous on the basis of just one input sequence (or even 1000 input sequences).1 (d) Let ¡ ¤ ¡ ¤ ¡ ¤ ¡ + ¤ Y + Y Y + Y Y ¡ ¤ + Y + Y EXAMPLE 3. converges to by part (a). that the use of deﬁnitions can be effective in demonstrating that a property (such as continuity) does not hold. then the output sequence would have been . so that for this input sequence. The conclusion is that ¡ ¤ § ¤ § ] is continu .
However. since this point is not in the natural domain of the function. is continuous on the open interval % r . For every sequence in converging to . For every sequence in converging to . Such functions illustrate the idea of continuity on an interval. E  C + C % ¤ r  Y ¤ + so that 4. Thus polynomials are continuous functions. % r + % r  + Y ¡  ¤ + Y % r ¡ ¤ 3.  C C C B B d + to be . we can write when B C C k It is not hard to see that every polynomial function is continuous at every point of its domain.15 is not continuous at the points and 1. since these domain of the function.9 The function C B B d 7 C C + r + B B B ¡ C 7 % C B d + + + + Y of Example 2. r + Y It should be clear how to deﬁne continuity on intervals such as or and we leave this as an exercise. EXAMPLE 3.CONTINUITY 37 EXAMPLE 3. DEFINITION 3.10 The function + + Y C B + Y C B is not continuous at . the sequence ¡  ¤ + Y ¡ ¤ 2. the sequence converges to converges to . Hence if we deﬁne is continuous everywhere.5 Continuity on a closed interval A function is continuous on the closed interval % r Y is said to be continuous on . This is because these points do not lie in the and . including the point . A function which is continuous at all points of its domain is simply referred to as being a continuous function . DEFINITION 3. A function which is continuous on an interval (closed. in its domain if: 1.4 Continuity on an open interval A function which is continuous at every point of an open interval It is also possible to deﬁne continuity on closed intervals. then converges to 4 for every sequence converging to 2. . open or neither) is said to be continuous on . No sequence of output values can converge to numbers do not exist.
2 In Exercises 1–6. the natural domain of the given function does not include the point each case. 5. assuming the pattern continues as indicated. the ﬁrst few terms of a sequence are given. up to a maximum of $24. 4. 23. deﬁne so that is continuous at . 25. (a) Sketch the graph of the cost of parking charges as a function of time.38 CONTINUITY AND SMOOTHNESS EXERCISES 3. In Exercises 13–20. decide whether the given sequence is convergent or divergent. 19. 9. Find an expression for the general term of each sequence. 10. 2. 20. If it is convergent. 14. 18. 3. " 2 Õ Ê 7 ¤ " ' ³ § g d f Õ f " 7 g  " E ³ 2 " É " " ° ' f d 7 d " f f 2 Ê d § Õ ¤ 7 2 2 " U 2 d  7 2 ' C ° " f " E 7 " d 7 7 " ¤ f Õ " § Ù " ' d ¯ 7 d E E " 7 ¤ 7 d " 2 2 É " ° ' f C C 7 7 C Ô Ô Ô Ô f " " " E " C g 7 f ° " Ü 7 ' " § f 7 " " E § 2 d " ° ³ 7 f 2 Ú Ö f f 2 f f ¤ f d " E 7 2 2 " Õ 2 " 7 " § ° C d " 7 ' 7 f d ¤ Õ C § E 7 E E E 7 d ¤ " f 2 2 2 d U g Ý " 7 7 C C C ¤ 7 f " C g v Ô Ô Û " 1. 16. 21. In Exercises 7–12. ﬁnd its limit. 22. 8. 17. . In 26. 24. 15. 7. 12. 6. ﬁnd the ﬁrst four terms of the sequence whose general term is given. 11. 13. In Exercises 21–25. The parking area in the Sydney Fish Markets charges $2 for the ﬁrst three hours or part thereof and $10 per hour after the third hour. where possible. (b) At what points is this function discontinuous? What is the signiﬁcance of these discontinuities to someone who uses the parking area? .
Let be the function deﬁned by the rule . 29. deﬁned by ß Þ 28. Show that the function 2 . 34. 7 r C E 7 r 30. . Draw the graph of each function and determine where the functions are continuous. Explain how we can deﬁne continuity of a function on intervals such as ¡ ¤ 2 or . E ) d " C Þ E ) C Þ E ) C à Ú ß Þ f E ) C ß á ß à Þ © à E ) C ß á à © h g h ä 7 g E r r for any . but for which . What does not converge to converge to? Find a sequence for which converges to 4. if if . Show that is such a sequence. Let be the Heaviside function deﬁned above. . . Show that the function deﬁned by is continuous at 2 E 2 C 7 ¡ ¤ 2 § 2 E 2 C 7 " f " Ô ¡ ¤ 2 33. Draw the graphs of . ﬁnd the corresponding sequence of for the functions indicated. It should be clear from your graph that is not continuous at verify this you need to ﬁnd a sequence which converges to 2. 37. ¡ E ¤ 2 C ¡ ~ " ¡ ¤ d 2 g ¡ ¤ 2 U E g C g g r w 2 2 ) 7 7 " " 7 d 2 2 v E ) C ß E 2 Þ C ) ß f C Þ 27. and the unit pulse function . Show that the function deﬁned by g 2 7 2 v E g 7 % C E l 7 " E " 7 l f C is continuous on each of the intervals and " " u 2 2 " f 2 2 v E 2 C g is not continuous at 38. and . Two important functions in science and engineering are the Heaviside function r u ) 7 % v . but not on the interval 2 2 g w 2 7 2 2 f d a r E 2 E C 2 C 36. Show that the function deﬁned by is continuous at a 2 2 d r E 2 C 35.CONTINUITY 39 for any . To does ¥ Þ . Õ ¡ E ¤ Õ 2 C For each sequence output numbers of input numbers in Exercises 31–35. deﬁned by ¡ E ~ g " ¡ E d ¤ 2 g 2 C C Draw the graph of . 2 d ' E 2 C 7 " f r Ô ¡ ¤ 2 2 d ' E 2 C 7 Õ " d r Ô ¡ ¤ 2 31. 32.
we introduced precise deﬁnitions for continuity and we must now do the same for differentiability. 4. In Chapter 3. In Chapter 2. The gradient of the tiability at the point .1 THE DERIVATIVE Let be a given continuous function and let be a point in its domain.1) ¥ P ¤ 2 ¤ 2 f ¤ 2 2 £ E E ¨ ¤ 2 2 C C .1: A secant of a graph ¤ 2 2 ¡ ¤ 2 Õ E 2 C 2 f f E ¤ ¤ 2 2 C E E Ô ¤ 2 C 7 ¤ If we take an input sequence secant joining and 2 C E E 2 C 7 2 C which converges to then for each point we can take a . we informally discussed continuity and differentiability. To investigate the differenin the domain of (Figure 4. we present a more detailed discussion of differentiability. The gradients of such secants form a sequence (4. In this chapter.1). ﬁrst take a point secant joining the points and is given by E E E 2 ¤ C 2 C f 7 E ¤ ¤ 2 C 2 C 2 2 k E ¤ E 2 2 C 7 2 C 2 ¥ ¤ £ Figure 4.CHAPTER 4 DIFFERENTIATION In Chapter 2. we informally deﬁned a function to be differentiable at a point if its graph has no corner at that point.
then it will converge to the gradient of the tangent at .25 5 2.1 12. we shall restrict ourselves to sequences as and for all positive integers . Such sequences will be called properties special sequences converging to .2: A nonsmooth graph i ¡ ~ " d g ¡ ¤ 2 g P 2 .1) may diverge for some (perhaps every) choice of 3 3 . then If several different things may happen: ¡ ¤ 2 The sequence (4. 2 å 2 2 ¡ ¤ 2 2 2 g 2 " k E ¤ 2 ¤ C 2 2 ¡ ¡ g ¤ ¤ 2 f 2 ¤ 2 2 2 2 ¡ å Ñ ¤ ) k Ï 2 Î Ð Í ¤ ¤ 2 g 2 ¨ 2 f ¤ 2 l C ~ ç E g 2 C f 2 E ¤ 2 C ¤ æ 2 2 .42 DIFFERENTIATION There are some minor sources of difﬁculty which we need to dispose of. In the ﬁrst instance.0661 1000 2. is in the domain of a function .24 10 2.1) for 1 ¤ 2 2 2.61 100 2. it appears that if the sequence (4. ¡ ¤ 2 ¡ ¤ 2 The sequence (4. we shall require that be deﬁned on In order for there to be sequences an open interval about . in which case the corresponding quotient in the sequence (4. we note that it is quite possible to choose a sequence which converges to and for which for at some positive integer .1) may converge to the same number for every possible choice of .001 12. . we have computed the sequence (4.1) may converge to a number which depends on the choice of 3 From a graphical point of view. .5 15.1 In the table.2 13.1) will which have the be undeﬁned.0067 3 19 E This table suggests (but does not prove) that E g C f E ¤ 2 C Figure 4. § and The sequence (4.01 12. To avoid this happening. and if is any special sequence converging to . EXAMPLE 4.1) of gradients of secants converges. with .
It is interesting to see how this corner manifests itself in terms of the gradients of secants. First take a decreasing input sequence for the case converges to . This is a function whose graph we would certainly Let us deﬁne the function by the rule expect to satisfy our intuitive notion of a smooth function. if with each and which converges to . if we take a sequence for which each and which converges to . let ¡ ¤ 2 and this sequence clearly converges to independently of the choice of the sequence that can be cancelled. then the diagram suggests that the gradients of the corresponding secants will converge to a negative number.2. that is. The output sequence (4.2) . Let % u 2 7 2 f v .THE DERIVATIVE 43 EXAMPLE 4. which has a corner at . We then have ¡ ¤ 2 Õ 2 f ¤ 2 2 Õ § E E 2 2 C 2 2 C { f E ¤ 2 C 2 2 2 2 2 2 u 2 ¤ ) w ¤ 2 ¡ ¤ ) ¡ ¤ 2 ¡ ~ " Õ .1 above. the two gradients will not then converge to the same number. In Example 4. Clearly. be any real number (so that is in the domain of ). because for every . On the other hand. In this example. so let us investigate the sequence (4. but such a table cannot give an exact answer for the gradient at a particular point. Notice . % k 2 f ¤ 2 2 g 2 f ¤ 2 Õ E 2 f ¤ 2 2 C f E % ¡ ¤ 2 2 2 2 d 2 f d ¤ ¤ 2 ¤ C 2 2 Ô Ô 2 E 2 C 2 f ¤ 2 Ô . we will use what we know about convergence of sequences to get an exact answer for the gradient of the tangent at a particular point on the graph.2 Consider the function whose graph is given in Figure 4. Choose any input sequence Let converging to . we estimated the gradient at a point on the graph of by drawing up a table.3 . Firstly. EXAMPLE 4. This input sequence  2  E 2 C EXAMPLE 4.4 Here is an example of a function which we do not expect to be smooth.2) corresponds in this case to E ¡ % C ¤ 2 f E ~ " C Õ E % C f E ¤ 2 C % % 2 7 % Õ f % % Õ ~ f f " E ~ ~ 2 " " C 2 Ô Ô f f E ¤ ¤ 2 2 C % Ô f ¤ 2 % Ô 2 The graph has a corner at . then the diagram suggests we take a sequence that the gradients of the corresponding secants will converge to a positive number.
while if there is the slope of the tangent at . verged to DEFINITION 4. + ð Y converging to . In this example the presence of the corner has caused different output sequences to converge to different numbers. the terms of the sequence (*) in Deﬁnition 4.44 DIFFERENTIATION which converges to . From a graphical point of view. The function if the sequence tiable at Y Õ + Y B B ¤ ¤ + Y Ô g Y E B is differen(*) g ¤ Notice that the statement that the sequence Õ + Y B ¤ + Y This deﬁnition says nothing about the interpretation of the value of the limit.1 get close to the for a function with no corner. where ¡ ~ B ¡ ¤ { which converges to . but the corresponding output sequence is Õ + Y B ~ B + Y Õ + Y B ¤ { + Y   % Õ B  ~ B + B Y Figure 4. as gets close to . then there is not a unique tangent. A function with no corner at a corner. On the other hand. We expect the limit of the sequence to be gradient of the tangent at has a unique tangent.1 Differentiable functions Let be a function which is deﬁned at all points of an open interval .3: The graph of    í + Y ë é B ~ ~ è B Ô Ô B { ¤ è é ë í è î é Ô î B ë ¤ î { í î Ô ï î also con for all . even though the corresponding input sequences both con. This limit is normally denoted by the symbol . ¤  + ð Y  + Y B B  ¤ ¤ + Y Ñ Ï Î Ð Í ¤  + ð Y converges to can be written as ¡ ¤ converges to the same limit for every special sequence . the increasing input sequence verges to .
The derivative of a differentiable function is one of its most important properties. it may also be differentiable.  + ð Y ¡ % % % is just the sequence which converges to . Thus h g % B B ¤ ¤  + Y B B  ¤ ¤ + Y ¡ ¤ .5 Suppose + Y Y ð Y ð ð Y + ð Y ][ ð Y Y Y Y Y ð Y Y + ð Y so that the sequence Õ + Y B B ¤ ¤ + Y Ô (4. however. It was discovered about 300 years ago by Sir Isaac Newton and Gottfried Leibniz. This method of solving problems by ﬁrst ﬁnding the derivative and then using this derivative to ﬁnd the solution function is the basis of calculus.2 The derivative The number is called the derivative of the function at the point . Let be a sequence converging to a point . we have to use the derivative to calculate the solution function. In this book. The deﬁnition we have given for the derivative of a function expresses precisely what we mean by a function having no corners in its graph.3. so that . we use differentiation by ﬁrst principles in the following two instances: To aid in the understanding of the deﬁnition ! ð Y EXAMPLE 4.3) E We leave you to prove this result as an exercise. There is a more important aspect to the derivative. Notice that since is a function. If this is the case. In many problems we ﬁnd that the natural formulation of the problem gives us information about the derivative of the solution function rather than information about the function itself.7 . Using the deﬁnition to ﬁnd derivatives is known as differentiation by ﬁrst principles . Then . The domain of is the set of all points where is differentiable. We can use this deﬁnition to calculate the derivative in a few simple cases. This enables us to deﬁne a new function (denoted by ) called the derivative of according to the rule . we denote its derivative by and call it the second derivative of . then according to the calculations in Example 4. then the derivative can often tell us a great deal about the behaviour of the function and may assist in sketching its graph. but it is not a very efﬁcient method compared with those which we shall shortly develop. except for the point If has domain and if if K M B v + ð Y ¡ Y B  h  + Y C  + Y ð EXAMPLE 4. . EXAMPLE 4. If we are given some function.THE DERIVATIVE 45 DEFINITION 4. fourth and higher derivatives.6 . If domain and is the function with domain given by the rule Y h ð Y + Y is differentiable at all points of its . then is differentiable at all points of its domain. where is any point in the domain of at which is differentiable. In the same way we can deﬁne third.
Leibniz wrote . We can regard such a letter as the name of the given function. 3 A process based on differentiation by ﬁrst principles also occurs in some numerical work. If the values of the independent variable are denoted by and those of the dependent variable by and if the function is deﬁned by a formula such as . You have probably met some of these before.46 DIFFERENTIATION To prove theorems which enable us to use more efﬁcient methods of calculating derivatives. We don’t even have to use letters. Thus we (incorrectly) speak of the function . To calculate the time rate of variation.1. we don’t have to restrict ourselves to single letters for function names. When measuring a variable quantity such as a seismic signal. Recall that the derivative of a function at a point is still used today—the { ' E 2 ' C { 2 { 7 E 2 C 2 f E 2 ¤ 2 C Ï Î Í E 2 C ð { E g C ð { ð 2 { ¨ ~ { ¨ 2 { l E 2 C 2 ¶ ~ { ¶ { 2 E 2 C f E ¤ 2 C { ¶ 2 ¡ f ¤ ¤ 2 2 where 2 ¶ is any special sequence that converges to . Particular values of are denoted by the function value notation. Common examples are . for example the square root function is denoted by the special symbol and the output number corresponding to the input number is denoted by . we need to consider some matters of notation. Of course. we often only refer to the formula in speaking of the function. and . 4. . then we write and say that is a . All of the functions we have discussed have been denoted by a single letter such as .1 Remarks on notation Before going on to develop efﬁcient rules for differentiation. Let us put and and considered the quotient " f " 2 g f 2 2 d f g ¤ 2 d 2 f 2 § f 2 § Ñ 2 Ð ¤ E 2 C E % " C % This means the same thing as If we want to substitute is " 2 if in the expression . Thus The derivative is written and this again denotes a function with particular values written with the function value notation. for instance. There is also an alternative notation for derivatives which originated with Leibniz and which is notation. function of . we may only have its values at discrete points in time. usually written as . We could. for example polynomials such as 2 ö õ Í 2 2 ò õ ô 2 ó Î ò 2 " f 2 E g 2 C d ö õ Í 2 f § E 2 2 C ò õ E ô 2 C E 2 C ó Î ò Indeed. we have to approximate the derivative by the slope of a chord joining two nearby data points. As this ( v . as in . when we really mean the function whose output number is when the input number is . Mathematica uses a similar notation. then the Mathematica expression . deﬁne a function ‘step’ by the rule % % w ) ) 7 7 " % E ) C step Some functions have names which are universally agreed upon. and . or . Many simple functions do not have universally agreed names and are deﬁned by algebraic formulas.x>10 Another notation is often used in applications where there are independent and dependent variables. This is the basis of ﬁnite difference methods. Function values in such cases are denoted by a vertical stroke notation: " f 2 g d 2 f § 2 " ÷  E " f 2 g d 2 f § 2 C " 2 f 2 g d 2 f § 2 xˆ3xˆ2+2x1/.
` § ÷  7 2 g § § ÷ ÷   2 2 2 ¨ ¨ E g C ð is the same thing as . but instead of writing for the derivative. It is . differentiate the given function by using the deﬁnition of a derivative. As simple examples we have ÷  E " f 2 g ú ú or d 2 f § 2 C 2 ¨ ¨ " f 2 g d 2 f § 2 E 2 C and this denotes the same function as if we had put the derivative are written with a vertical stroke. Leibniz wrote and . 8. then if if " d 2 U d § 2 E 2 C ( " d 2 g E 2 C 11. where there is no suggestion of division.1 E ' C ð In Exercises 1–6. E g d 2 C ~ " E 2 C d 2 E 2 C 3. Function values of 2 ¨ { ¨ { 2 ¨ ¨ 2 { called it the derivative of with respect to . In each case. " In Exercises 7–12. " d 2 E 2 C ' E 2 C 7. 12. the . 4. plot the graph of the derivative. 10. At what points is differentiable? % % u w 2 2 " f " v E 2 C ð  2  E 2 C 13. ﬁnd " d 2 U E 2 C by using the deﬁnition of a derivative. In the Leibniz notation. E g d 2 C ~ " E 2 C g f 2 ' E 2 C 9. 2 d § 2 1. 2.THE DERIVATIVE 47 would be written as or E " f 2 g d 2 f § 2 C 2 ¨ ¨ { 2 ¨ ¨ EXERCISES 4. Thus { ¨ ð " f 2 g d ÷ ú ú 2 2 f ¨ § 2 { ð just another way of writing derivative of the function . Show that if . 2 ~ " E ' d 2 g C E 2 C 5. Note that does not mean divided by { 2 ¨ ¨ E 2 C ð E 2 C ð quotient approaches . ð ð Plot the graph of . 6.
{ 16. For instance. Decide which. As yet. { 17. except by the use of Deﬁnition 4. % 2 and is differentiable at 2 2 C . of the two functions . justify your answers. The functions we have already met have often been combinations of simpler functions. 2 4. the function E g d 2 C d § E " d 2 C E 2 C 2 E 2 2 C (b) is the population of a city at time . In this section we will develop more efﬁcient computational procedures for evaluating derivatives. which is often a cumbersome process. draw the graph of the derivative of the given function. if either. 2 E 2 C (a) (c) is the height of a missile E seconds after liftoff.1 on page 44.2 RULES FOR DIFFERENTIATION We have identiﬁed the derivative of a differentiable function as one of its important properties. E 2 C f E ( ( d ( 19. is the cost in dollars of producing units of a product. Let be a differentiable function and let to the quotient 2 C be a small positive number. { in the following cases. { 18. Give a physical meaning  § 2  E 2 C 2 2 (  2  E 2 C 14. we have no way of calculating the derivative. 15.48 DIFFERENTIATION In Exercises 15–18.
1 The derivative of a constant function Suppose is a constant function given by for all and some real number . while the function and be a . DEFINITION 4. THEOREM 4.1 on page 34. and An alternative and traditional way of writing the above result is to put Then 0 E 2 C ü l 0 2 ¨ ¨ d ü 2 ¨ ¨ E 2 C ð E 0 d ü C 2 ¨ ¨ E 2 C ð d E 2 C ð E 2 C ð E d C ¡ ¤ 2 We prove this result by using Deﬁnition 4. The domain of is the set of all points where and are both rule deﬁned.2 The derivative of a sum Suppose and are both differentiable at a point . products and quotients of given functions. Then . The sum of and is another function. . EXAMPLE 4. One of the most common ways of forming new functions is by using sums. so let us now be precise about such combinations of functions and their derivatives.9 As a very simple example we have i " d 2 g E 2 d 2 C ¨ E E 2 2 C C ð The ﬁrst term on the right hand side approaches as . deﬁned by the Let and be functions. Then E g d 2 C E g E d 2 C 2 C § E " d § E 2 C " d E 2 C 2 C û E 2 C is the sum of the two functions and . we can differentiate quite complicated functions. Then E E 2 C d E 2 C C f E ¤ 2 C d E ¤ 2 C E 2 C E d C f E ¤ 2 C E 2 d C 2 E 2 C 2 f f E ¤ ¤ 2 2 C 2 f d d ¤ E THEOREM 4.8 Let g d 2 E 2 C d d E 2 C d E 2 C E 2 C E d C ± 2 ± E 2 C û 2 % E 2 C ð EXAMPLE 4. . This means that with a relatively small number of derivatives and rules for the differentiation of combinations of functions. denoted by . while the second approaches Thus . The proof is left as an exercise. Then for all . It turns out that we can use the derivatives of and to ﬁnd the derivatives of and . 2 E 2 C ð d E 2 C ð E 2 C ð E d C is also differentiable at E " d 2 g 2 " 2 d C d § 2 2 2 C f f g E E ¤ d 2 d ¤ 2 C E 2 2 C 2 g " ' d d E d d 2 C § 2 C 2 2 g 2 d ¨ E § 2 C 2 E 2 E f d 2 C ¤ C 2 and let 2 .RULES FOR DIFFERENTIATION 49 is the product of and .3 Sums of functions .1 on page 44 and Theorem 3. Let sequence converging to .
! The derivative of a scalar multiple of a function is the same scalar multiple of the derivative of the function. ! Unfortunately. then is the function deﬁned by + Y ± + Y ± + Y ± is any real number and is any given Y  + Y Ê Y Y É Y Y . for any real number . we have Y h Y is also . so that the domain of As for the quotient of these functions. It is easy to verify that. + ð + Y d + + ð Y + ð Y Y + is also differentiable at E ¡ % B h The domain of the quotient function is .50 DIFFERENTIATION and you should do this as an exercise. + + Y + Y + The product of and is the function deﬁned by the rule product is deﬁned at all points where and are both deﬁned. h B Y The domain of both and is . We can summarise the two rules we have so far: The derivative of a sum is equal to the sum of the derivatives.3 The derivative of a product Suppose and are both differentiable at a point . + ð Y ± ± + Y + ð Y ± + Y 7 [] then is the function given by the rule Y ±  We can also deﬁne scalar multiples of functions: if function.4 Products and quotients of functions Let and be two functions. Then the product and .  B +  +  Ê +  É Y + B  + EXAMPLE 4.10 Let and let + Y . The ! . . the results for products and quotients do not follow such a simple form. Y Y The quotient of ! and is the function deﬁned by the rule Y THEOREM 4. Then k  + Y The quotient is deﬁned for all points where and are both deﬁned and where Y 7    + + It has the same domain as . For example if . First a deﬁnition: DEFINITION 4.
Note that the derivative of the product is not the product of the derivatives.4) in terms of E 2 C f E ¤ 2 C and 7 E 2 C 2 f f E ¤ ¤ 2 2 C since these terms converge to and respectively. There is a trick used here: subtract and then add the term in the numerator of equation (4. In fact E ¤ 2 C 2 2 E 2 C d E 2 f ¤ 2 2 C E 2 C f E ¤ 2 C l E ¤ 2 2 C E ¤ ¤ 2 2 C E 2 C Theorem 4. we take a point in the domain of and an input sequence to . E 2 C 0 E 2 C ü d i E 2 C E 2 C ð where we have used Theorem 3.4).4) E 2 C E 2 C 2 f f E 2 ¤ 2 ¤ 2 f 2 f C ¤ ¤ 2 2 d E ¤ 2 2 C f E ¤ 2 C 2 2 f f E ¤ ¤ 2 2 C 2 f ¤ 2 2 f ¤ 2 (4. but how can we assume that is continuous at ? Well. Then we must consider the sequence whose general term is E 2 C E 2 C f E ¤ 2 C E ¤ 2 C E 2 C E 2 C f E ¤ 2 C E C 2 converging (4.3 can be written in Leibniz notation as 7 ü ¨ 0 d 0 ¨ ü E 0 ü C ¨ where and . then apas approaches .5) approaches and this is the derivative of . as given by Theorem 4. ¡ ¤ 2 To justify this result. E 2 C E C E 2 C ð E 2 C E 2 2 C ¨ ð E 2 C 2 E ¨ 2 d C 2 % 2 f § f f ¤ E 7 E ¤ E 2 2 ¤ C 2 2 ð 2 C C 2 ¨ l Now as . Thus the right hand side of equation (4. This gives E 2 C E 2 C f E ¤ 2 C E 2 C d E ¤ 2 C E 2 C f E ¤ E 2 2 C C ð E ¤ 2 C E 2 C ð E 2 C E E ¤ 2 C C f E 2 E C ¤ 2 C E C and 7 E 2 C ð E 2 C 2 f f E ¤ ¤ 2 2 C but what does approach as ? We know that if is continuous at .3 is an inevitable consequence of the deﬁnition of the derivative and its interpretation as the slope of the tangent or (as we shall see in the next section) as a rate of change.1. The correct rule.RULES FOR DIFFERENTIATION 51 The aim of the game is to write the left hand side of equation (4.5) . it is true proaches in general that if a function is differentiable at a point then it is automatically continuous there.
we state the rule for differentiating a quotient. but we shall not present the proof here.11 We have ¨ ~ % k E 2 C E ¨ the same result we obtained earlier by ﬁrst principles.52 DIFFERENTIATION EXAMPLE 4. We can do this because we have proved some theorems. We had to use the original deﬁnition in the proofs of the theorems. The purpose of theorems is to simplify calculations. Next we have ¨ ¨ and in general i ¥ ¤ 2 E ¤ 2 C 2 ¨ ¨ The last result in the above example is an important one and is worth highlighting. we can differentiate any function deﬁned by an algebraic formula without having to go back to the original deﬁnition.2. . With the results in this section.4 If is a positive integer and There are a number of ways of proving this result. then we would have to go back to the deﬁnition to ﬁnd its derivative.5 The derivative of a quotient Suppose and are both differentiable at a point is also differentiable at and 2 E 2 C ð ð 2 and suppose This result can be written in the Leibniz notation as 7 ð 0 ü f ð ü 0 ü ¨ where and . then 7 7 2 2 ' g 2 2 " 2 g 2 ¨ 2 § ¨ ¨ § E ¨ E 2 2 2 2 2 2 d § d d § d 2 " 2 2 " C 2 ¨ 2 2 § ¨ ¨ C 2 § ¨ ¨ 2 2 2 ¨ 2 2 þ E 2 E E ¨ 2 2 § C C 2 2 C 2 Ê ¨ 2 E ¨ 2 É C . but once we have proved the theorems we no longer have to differentiate by ﬁrst principles. Finally. One way has been suggested in the above example and another method is outlined in Exercise 4. This rule can be proved along the same lines as for a product. THEOREM 4. THEOREM 4. if we encounter a new function which is not simply a combination of known ones. Then the quotient . Of course. E 2 C 0 E 2 C ü ¥ ¤ 2 2 C ð E E 2 2 C C ð E 2 f C 0 E 2 C h g ÿ 0 2 for ¤ .
2 In the following Exercises. ACCELERATION AND RATES OF CHANGE 53 EXAMPLE 4. where is some constant. in the period 1665–1667. 2 E ¨ E 2 C " E 2 f g E C " 2 C f f E 2 2 2 C C 2 g f E 2 C ð . Show that ¤ § ¥ ¤ ¥ ¤ E 2 C r 2 f E 2 C be an arbitrary sequence E ¥ ¤ r d ¥ ¥ ¤ ¤ 2 r d 2 d E § 2 ¥ ¦ C 2 ¤ ð ¦ ¦ r d d ¦ 2 § ¦ ¥ ¦ d ¤ ¦ ¤ ¦ ¦ 2 r d 8. then .3 VELOCITY.12 Let " f 2 2 2 C ð E ¨ E ¤ " 2 ¨ " E 2 ¥ 2 E 2 ¥ C 2 C 2 f 2 ¤ " C ± C f ¤ g 2 Then E " f 2 C ¨ 2 f E 2 C f 2 C EXERCISES 4. laid the foundations of the calculus. the derivative ﬁrst explicitly appeared in connection with the motion of bodies. then .5 to show that if ¥ ¤ 2 converges to . Show that if for all . where is a negative integer. E " f 2 C E " d 2 g C E 2 C 2 " d " E 2 C 3. let which converges to . 7 ¥ ¤ 2 d ¡ 2 2 ¥ for any real numbers and and any positive integer . 6. This was an innovation by Sir Isaac Newton who. " d 2 ' d 2 g E 2 C 2 ' d 2 E 2 C 1. 4. where % E 2 C ð d 2 d 2 2 ± 2 f E 2 C 2 7. 2. Begin by checking that r d ¥ ¤ C E r f C ¤ r f ¤ . differentiate the given function using the theorems in Section 4. Use Theorem 4. Here is an outline of a proof that if integer.VELOCITY.2. ¥ ¤ 2 E E 2 C 9. Next. ACCELERATION AND RATES OF CHANGE Historically. is a positive E " d 2 g C E 2 d ³ 2 U C E 2 C g " d f 2 2 E 2 C 5. then . . § and hence that the sequence Õ E 2 C 2 f f E Ô 4.
ﬁnd the rate of change of the area of a given square when its side length is 6 cm. then decreases as increases. First take a special sequence any particular value of .3.3. 4. If is increasing at . . while if it is positive. particularly the motion of the planets around the sun. The sign of is signiﬁcant. where is a smooth function. that is.54 DIFFERENTIATION Newton was interested in the mechanics of motion. which he called ﬂuxions. Consequently. If the length of side of each square increases in length with a speed of 2 cm/sec. which converges to . we can proceed in a similar way In the more general case of a differentiable function to which we did in the case of velocities. the velocity is given by ) ) ¤ ) ¡ ¤ ) ) ) f ¤ ) Ô E ) C º ) . so that at the time . The number to which it converges is deﬁned to be the instantaneous velocity at the time . EXAMPLE 4. the average velocity between the time and the time is given as average velocity distance moved time taken E ) C ) f f E ¤ ) ¤ ) C As we let get arbitrarily large. Solution. . If changes by an amount much.2 RATES OF CHANGE { ¶ ¶ ¤ d 2 { . If it is negative. His notation of for is still widely used today. For We can do this in the following way. the slope of the line gives the rate of change of with In the case of a straight line then the corresponding change in is times as respect to . The area of a given square is given by 2 £ £ 2 { 2 { { { { E 2 C { 2 2 2 { { E 2 2 C ð ¶ { 2 2 2 ¶ 2 2 ) ! E ) C ð £ ) ) g 2 since at time . velocity is the rate of change of distance with respect to time. say. We deﬁne the instantaneous rate of change of with respect to at the point by . In general the velocity at which the particle time is given by moves will vary as the time changes and we may wish to ﬁnd the velocity at a particular time . But the number to which the sequence Õ ) E f ) C ¤ ) f E ¤ ) C ) 4. then the rate of change is positive. the rate of change is negative. while if is decreasing at .13 A screen saver consists of expanding squares which appear at random points on a computer screen. He restricted his attention to derivatives with respect to time. which involved velocity and acceleration. the time interval will decrease and we expect that the sequence of average velocities will converge.1 VELOCITY » º ) ¨ ~ 2 ¨ £ 2 Suppose a particle moves along a straight line in such a way that its distance from a ﬁxed point at . then increases as increases. As an example. 0 E 7 ) ) C U ð 0 ) E ) C ð converges is just .
what is its velocity after 4 seconds? 6. ) By using a similar argument to the above. This causes a circular ripple that moves out from the point of impact at 2 m/s. Find the average velocity of the body over the time interval and . Use the results from Exercise 2 to guess the instantaneous velocity at ( d ' 7 ' ( d ) 7 ) 1. ACCELERATION AND RATES OF CHANGE 55 4. Tabulate the average velocities over the time intervals . ) f ) 2 axis in such a way that its position at time is . we deﬁne the instantaneous acceleration at a time r 0 ¨ E ) C ð ð ) U § f g ¦ ú ú ú ) ) ¨ E ' C ð r ` Since when time is 24 cm /sec. we compute £ .3 ACCELERATION ) 0 Finally we consider the concept of acceleration.3. What is this rate of change when the radius is 3 m? 2 8. At what rate (in m /s) is the area within the circle increasing when ? 10. r ) ' ) ( 2. The distance of a particle from a ﬁxed point at time is given by ) ) ) r " d ) f ) § ) U º º l 7 7 ' 7 g 7 " ¤ ( d ' 7 ' ¡ ¤ ( At what times is the velocity zero? What is the acceleration at these times? 7.3 ) 2 ú Exercises 1–4 refer to a body which moves along the given by . Show that the rate of change of area of a circle with respect to its radius is equal to its circumference. 9. Find the rate of change of the area of a circle with respect to the radius. for seconds. Find the rate of change of the area of a square with respect to the length ? What is the rate of change when ) % ) ' 2 of one of its diagonals. Conﬁrm your guess in Exercise 3 by ﬁnding the limit of the average velocities over the time intervals . Show that the rate of change of volume of a sphere with respect to its radius is equal to its surface area. Hence the rate of change of area with respect to to be . 3. . 5. Let be any sequence converging to zero. " % % % " % % 7 " % 7 " 4. A pebble is dropped into a pond at time . EXERCISES 4. If a particle has a velocity at time and a velocity at time . where is measured in seconds. 2 ' ) . then the average acceleration during the time interval is deﬁned to be ) f ) 0 f 0 ) 0 The acceleration is the rate of change of velocity with respect to time.VELOCITY. 11. If an object has fallen a distance of metres after seconds.
2 4. Melbourne. 0 1 º ) 2 3 4 5 6 7 8 9 10 0 1. What is a reasonable domain for ? What can you say about the sign of £ º ) ` r ) r U ) ' ' g " º 11 12 ` r r % E ) " " % C % % % ) for ? . What is the physical meaning of this result? © E º C ð 14. When could this person legally drive home? [From Petocz.7 10.0 Find the average velocity during the time interval . Newton’s law of gravitation states that when a spaceship is at a distance from the centre of the earth.3 30.8 140.6 19.4 162. Over which period of time is the car moving with constant velocity? ation at E º C 15.2 58. is the mass of the spaceship and is a constant.56 DIFFERENTIATION 12. Estimate the velocity and accelers.0 43.6 75. Petocz and Wood. The distance (in metres) of a car from a ﬁxed point at time (in seconds) is given in the following table. 13. known as the gravitational constant. Thomas Nelson. The following graph shows the percentage of alcohol in the blood of a person as a function of time in hours since drinking two double whiskies.6 97. The elevation of the Parramatta River km from its source near Parramatta is given by some function .05%. What was the approximate rate of increase in the percentage of alcohol in the blood after half an hour? At approximately what rate was the person eliminating alcohol after four hours? The legal blood limit in most states in Australia is 0.] £ E ) C ) 16. 1992.2 118. its weight is given by 7 © © where is the mass of the earth. Introductory Mathematics. Show that the rate of change of volume of a cube with respect to the length of its edge is equal to half its surface area. Find the rate of change of with respect to and show that it is negative.
that is. as in the Tower of Terror. we examine the motion of bodies moving under the inﬂuence of gravity. The rule deﬁning will ( E ) C ( ) g " d d 2 2 ' f d 2 2 E 2 C Rational functions. We are seeking a function such that . which are quotients of polynomials. So far the only functions we have considered are those deﬁned by algebraic formulas. which in the case of a rational function is the set of all points for which the denominator is not zero. Thus in searching for a solution to a practical problem. For example or Such functions are differentiable at all points of their domain. ° ¥ 2 E 2 C .1 THE TOWER OF TERROR Suppose. f § 2 E 2 C " d 2 ' d 2 g . a function which is continuous and differentiable. that of falling bodies. For example E 2 C Polynomials. In this chapter. 5. For example or . We have already set down the requirement that the solution to any of the problems we will consider in this book should be a smooth function. where is an integer. the ﬁrst place we might look is among functions of the above types. which are the sums of positive integer power functions. Let be the height above the ground and let be the time.CHAPTER 5 FALLING BODIES We now have enough mathematical machinery available to solve the ﬁrst of the problems posed in Chapter 1. Here are a few of the cases where the mathematics needed for this type of problem is used: In the Tower of Terror described in Chapter 1 3 3 3 3 3 3 3 In the motion of rockets and missiles In designing artillery In some movie stunts. stop at its highest point and then fall. we project a body upwards from some point on or above the Earth’s surface and observe it rise. There are three main types: 2 E 2 C ¤ 2 E 2 C r Power functions of the form .
what is correct and what is convenient may sometimes be two different things. The problem is now to ﬁnd some way of predicting values of for different values of . as in “by abuse of notation. Notice the role played by the various symbols: is a number which is calculated from another number by the application of a rule . Mathematicians have introduced special terminology to describe such cases. people have been hurling objects through the air.1 to read off values of for values of other than those in Table 5.4 9 14. The symbol is thus doing double duty.4 3 102. In practice.9 4 112. the computational rule. However.1: Hypothetical data values for the Tower of Terror There are at least two ways to proceed: By using graphical techniques.1). graphical techniques are applicable to other problems where the theory may not be well understood.9 8 53. We can use the curve in Figure 5.1 15. We begin with a set of hypothetical experimental results that could apply to the Tower of Terror (Table 5. starting when the car is at the beginning of the vertical part of the tower.1. This may be as a result of some theoretical considerations or it may be purely empirical. Our aim is to ﬁnd the function. that is. The theory is now well understood. However. ! By developing a formula which expresses the height as a function of time. where friction has to be taken into account. apply to a body which has been ﬁred into the air. but this was not always the case. The motion will thus be quicker than that of the car in the Tower of Terror. It was regarded as important to understand matters such as the ﬂight of arrows or cannonballs.6 7 82. we would probably not take such an approach for this problem. They call it abuse of notation . ! We shall explore the graphical method ﬁrst. which enables values of to be calculated from values of . Throughout recorded history. In what follows. We shall ignore all questions of experimental error. you will learn methods that will be applicable to more complex problems.1. in fact.1 2 83. There were many misconceptions about the laws governing the motion of projectiles and it was not until the 17th century that Galileo ﬁnally developed the theory that we use today. Since we now know how to predict the path of projectiles we could simply tell you the relevant functions.58 FALLING BODIES tell us how to calculate values of from values of . In the righthand diagram we have drawn a smooth curve through plot of the points these points. usually at each other. It may denote either a number or the rule for calculating this number. is the value of when ? 0 1 54. measured in seconds. What. These ﬁgures. This procedure is not entirely satisfactory for a number of reasons: % +  + Y  + Y  + 7  + Y . because the mathematics and physics needed to ﬁnd solutions are well understood. This is what we have to do if we wish to keep things strictly correct. We shall follow an idealised development of the problem and omit the errors and blind alleys that actually occurred in the development of the theory.0 Table 5. that is. We measure the height in metres above the ground at a sequence of times . we use trial and error to ﬁnd a formula that ﬁts the known data and use it to calculate any unknown values. The diagram on the lefthand side of Figure 5. let us write The problem we are considering is a very old one.5 6 102. for ”.6 5 112.1 shows a in Table 5. this is only one of the many practical problems to which mathematics can be applied. However. By looking at the development of a simple problem such as this one. it will be more convenient to write instead of . for example.
the slope of the chord joining two points gives the value of the average velocity between the two times and the slope of the tangent gives the instantaneous velocity. from Figure 5. we would have to start again from a series of measurements. Table 5.2 gives the velocity for various times . The velocity can be calculated by estimating the slope of the tangent at a number of points. even if the graph is in some sense correct. Obviously. ! The results only apply to this particular case. One possibility for such a quantity is the velocity at different times. See. ! We are making assumptions about the behaviour of the curve in regions where we have no data. where the velocity is estimated to the nearest 1 m/s. ! Let us now try to ﬁnd a formula for the height in terms of the elapsed time. A faster initial velocity should result in a greater maximum height. then we can calculate the velocity by differentiation.1: Hypothetical data values for the Tower of Terror We have to make some assumptions about the general shape of the interpolating curve. Figure 3.1 with straight lines? ! There is an inherent inaccuracy in reading ﬁgures from a graph. For example. In the present problem. For instance. it would be preferable to have a method of calculation that relies on only a few input parameters. if we know the equation of the we can calculate the velocity at curve. why don’t we simply join the dots in Figure 5. It would help considerably if we could ﬁnd some other quantity to plot that would give a straight line. but in this case it is the equation that we are trying to ﬁnd. H Á < 7 C . This may or may not be useful. It is easy to ﬁnd a function represented by a graph when the graph is a straight line. for example. which we would naturally expect to be an important factor. such as the initial velocity.4 on page 29.THE TOWER OF TERROR 59 Figure 5. As an experiment we try plotting the instantaneous velocity against time. but not in other cases. If we wished to calculate height values for some other projectile. Of course.2 to be about m/s.
C < 7 Á < 7 C .
C < 7 .
we have plotted these points on a graph. The points lie almost exactly on a straight line. we can make use of the Mathematica command Fit which will derive the parameters for the best straight line through two or more plot points.3: The velocity of the projectile ) ( Recall that we are seeking to express the height equation 0 ( ¨ in terms of .3.2: The velocity as the slope of the tangent (seconds) ) 1 35 2 24 3 15 4 f 5 f r Table 5. If we wish to be more accurate. We draw what appears to be the best straight line through the plotted points and then ﬁnd its and the slope is found to be . Since the velocity is given by the U ' f U g f U " (m/s) 0 5 Á Á ! 7 ! E ` ) C ` U ) ¨ U g g % % % % % g g U U f f . The intercept on the –axis is . ! f U U 0 ) ! f U U E ) C 0 Figure 5. so the equation is equation.60 FALLING BODIES 44 3 % Figure 5.2: Velocity at different times In Figure 5.
% % % % g % ! ` U " 6 7 8 .
For values connecting the dependent variable with the independent variable .THE TOWER OF TERROR 61 we have (5. 1. 3. There is also a numerical method of approximating derivatives from a given table of values. 4.8 3 1. The problem is now to solve the differential equation we have obtained. and not a number. In obtaining Table 5. 2. let . Thus next point ' ) ! " U " f g ' E ' C ð E g ' 7 U C Central differences take the average of the slope of the chord on either side of the point in question and are usually more accurate. This is . that is.4 4 3.1).1) This is an equation containing the derivative of the function we need. Equations such as (5. Note carefully that the unknown quantity in equation (5.1) is a function. we had to graphically estimate the values of the derivative.2 from Table 5. 5. 0 E ) C ) 1 0. where the unknown is a function and the derivatives of the unknown function appear. Calculate the derivative of E 2 C ð { (c) The method of central differences approximates the derivative { f § { E 2 C ð { by E 2 C ð { (b) The method of backward differences approximates the derivative { f { E 2 C ð { by E § { 7 ( § 2 C 2 E ( { 7 2 C 2 2 2 E ) f f f ! { 7 § § 2 2 2 2 f C U U § ( E 2 ) C ð u { ( 2 u 2 at by using to the . and be three tabulated the method of differences . Thus g ' " ) ! % f g ' E ' C ð E g ' 7 U C ' ) g ' f f U U % ) EXAMPLE 5.2 2 0.1 A quantity is measured at times . E ! % 7 g C (b) The method of central differences takes the slope of the line joining the point point as an approximation to the derivative at . are called differential equations . Then E 2 C ð { (a) The method of forward differences approximates the derivative { f § { E 2 C ð { by (a) the method of forward differences (b) the method of central differences.1. The function is called the solution of the differential equation. We will address this problem in the next section.2 5 5 to the E U " 7 ' C 0 (a) The method of forward differences takes the slope of the line joining the point as an approximation to the derivative at . Differential equations are at the heart of the applications of the calculus and we shall encounter many of them. we have to ﬁnd the function . and we have to use this equation to ﬁnd the function.
using your approximate results in part (a). Tests on the C5 Chevy Corvette gave the results in the table below. Let us try this for the equation (5. (b) Plot the graph of above. This is a polynomial of degree 1 and. We have ) ! ¤ d ! f ) U U d f U ) E U r ¤ d ) E ) C ( d ) r C ¨ that is. 0. . how fast would it ) C (c) Use your graph to write down a differential equation for . for r % 7 " % .175 ) 0. but is really just another name for intelligent guessing.112 ) C ð 0.292 0. We want to ﬁnd an expression whose derivative is since differentiation reduces the degree of a polynomial by 1.0 0.4. % ) E 7 2.9 0.3 0.247 0.1 (a) Use the table to calculate E ) C ð approximately. You are conducting an experiment in which you measure a quantity 0 $ 4 " ! 4 at different times . This is referred to as using a trial solution .148 0.103 0.343 1. for the whole period. (m/s) 7 ` 7 ' ) (a) Estimate the acceleration for 5.4 0.2). 0. (5. Check your formula for the case 0 0 3 20 6 33 9 43 . over the period shown? Justify your answer.127 E 0.62 FALLING BODIES EXERCISES 5.0.208 " % 0.1 ) 1.5.1 0.5 0. 0. ) Time. on the interval .6 0. then the coefﬁcients ) ' ) (c) If the car were able to maintain its acceleration at be going after 12 sec? E ) C ð 0 (b) What can you say about the sign of 0 ) ¨ E ) C (d) Deduce a formula for . it seems reasonable to take a polynomial as the trial solution to substitute into of degree 2 as a trial solution. .2 SOLVING DIFFERENTIAL EQUATIONS One general method of solving a differential equation is to make a reasonable guess as to the form of the solution and then substitute this guess into the original equation. Velocity.8 0.2.4 0.2) ) ) ! f U U E ) C ð ( .3. Here we are making use of the result that if two polynomials are equal for all values of .7 0. 12 51 . 0. Let us take equation (5.2 0.3) U U U f r ) ! f U U d ) r g Comparing the two sides of this equation enables us to deduce that and .
if then H d H § 7 7 d H § 7 B H + H d 7 7 d % 7 H B C  + H ¤ so that and the solution to the above initial value problem is % H ÷ ¦ ú ú Á ç  ¤ + d 7 7 d < 7 7 B % æ ¤ C C < 7 . This information. Hence (5. together with .1 if our analysis is correct.SOLVING DIFFERENTIAL EQUATIONS 63 of corresponding powers are equal. we cannot get the value of this constant from the differential equation. equation (5. . it is the initial height which determines which curve to select for a given case.4). Unfortunately. The reason for this is that the information we get from equation (5.4: Solutions for different initial heights H + In the present case when . that is.4) There is still the constant to ﬁnd before we know .2) can be written as an initial value problem % B 7 7 + ð H where we give both a differential equation and a value of the dependent variable corresponding to a given value of the independent variable. equation if we had started the motion at a height of say C + Y ¤ ¤ d 7 7 d 7 B + Figure 5. From equation (5.4). As we shall see in a moment. We would have obtained the same differential m or m or m. For instance. we have + H H This equation should reproduce the values in Table 5. The justiﬁcation for this procedure is provided by Theorem 5.2) is the slope of the tangent to the graph of at the point and there are many different curves with the same slope (Figure 5.1 on page 64.
1.64 FALLING BODIES which agrees with the result in Table 5.2 7 ¤ 7 7 r In Exercises 1–4. % ¤ d 2 d § 2 r 1. and . then 2 for each .3 Solve the initial value problem { ¨ EXERCISES 5. Thus the solution is 7 ¨ d ) U d ) g ' d § ) " ' { g ~ ' ' ~ " r so that . It may be that if we were to measure the height for a value of other than one in Table 5. 4. . However.3). and c=4. we might ﬁnd that equation (5. We leave you to check the rest of Table 5.2). we have made use of the following result. Remark In ﬁnding the values of and in equation (5. and .2) did not agree with the experimental result. " " d ) U ¨ d ) g ' % d § ) ) " ' { " { and using the initial condition when gives .1. r ¨ U ¤ g ' r ) r d ) U d ) d § ) ' ¨ d ) ¤ d ) EXAMPLE 5. .1 Equating coefﬁcients If the two polynomials ¤ r d 2 ¥ ¤ r d ¦ ¦ ¦ d ¥ ¤ 2 r d ¤ 2 r E 2 C 1 ) r EXAMPLE 5. then 7 7 g 7 " 7 % 2 ) ) r are equal for all values of . Of course. until this happens (and it never has for falling bodies).2 If g d § ) r 3 for all . this is not conclusive evidence that we have found the right function. we retain our solution (5. THEOREM 5. determine the coefﬁcients ¨ d 2 ' d 2 2 ¤ d 2 d § 2 r r d ) g d ) ' ¤ d ) g d ) r ¤ d ) ' d § ) ) ' d ) ' d § ) r 3.1 as an exercise. Thus 7 U d ) ' d ¨ ) d ) ¤ ¤ d d ) ) g d d § ) r ) r ' { Solution. This would indicate that there was an error in our theory and we would have to reconsider the situation. We take a trial solution gives and substitute it into the given equation. This r ¨ U " ¤ g E % C { % 7 U d r ) ' d r ) d ) U ) ¨ d ) ¨ d ) ¤ d ) g d § ) r If 3 then . ¤ d 2 ¥ ¤ d ¦ ¦ ¦ d ¥ ¤ 2 d ¤ 2 E 2 C ( and . 2.
14 2. observers have made errors.46 1. ' E % C { 7 " d ) { ¨ 7. A new set of experimental results would require us to repeat the above rather lengthy process. Solve the initial value problem. 0. Each storey is 3 metres in height.1) only applies to the table of experimental results of Table 5. Solve the initial value problem.3 GENERAL REMARKS The procedure we have used in the previous section gives rise to a number of questions: We have obtained the equation (5.41 .24 1. % E % C ( 7 ' d ) g d § ) E ) C ð ( 8. Is there any way of getting a more accurate equation? 3 Equation (5.1) by experimental means. Justify your assertion. Can we get a general equation which can be easily applied to all situations? 3 ¤ r r ¨ r f d 9 7 7 9 d 7 d 9 ) ¨ 2. Observers with synchronised watches are stationed at windows and note the time at which the stone passes the midpoint of their window. They compare results and arrive at the following table of the time (in seconds) taken for the stone to reach the midpoint of successive windows.96 1.83 1. It is possible that one or more of the observers may have made an error. The windows are halfway between the ﬂoor and ceiling. 5.28 2. If show that f 9 d 7 d 7 E 9 f 2 C E 7 f 2 C E f 2 C r ¨ d 2 ¤ d 2 d § 2 r 6. A stone is dropped from a 10 storey building. so it is.GENERAL REMARKS 65 5. Check that your equation adequately reproduces the results in the table and decide which. a good approximation to reality.1. Solve the initial value problem.55 ) 0.66 1.91 Use the graphical techniques described in this chapter to determine the equation of motion. g E % C 0 7 U d ) g d ) E ) C ð 0 9. at best. if any.
Subsequently we can use these laws to ﬁnd the differential equation in any particular case. is the initial velocity. Since . we have another differential equation (5. although it is unclear whether or not this story is true. More accurate measurements give ms .7) corresponds to the experimental result of equation (5. projectiles describe a curved path of Galileo was also the ﬁrst person to realise that the path of a some sort: however. The physical law that Galileo discovered is the fact that all freely falling bodies have the same constant acceleration. Initially. whose derivative is . The second use is to investigate the properties of these functions.5) only gives the slope of the line in equation 5. In the present case this is not difﬁcult.6. The acceleration is given by . In fact.6) ¤ E ) C 0 d ) f E ) C 0 f f E ) C 0 is such a function. no one has pointed freely moving projectile is a parabola. In order to solve this differential equation. whereas is a function. The ﬁrst use is to ﬁnd functions which govern the variables in practical problems. commonly denoted by the symbol .5) f is the slope where is a positive constant whose value has to be determined by experiment. commonly denoted by the symbol . where is any constant. where We can now use our solution to compute various properties of the motion.8) . Galileo’s law states that the velocitytime graph for all bodies falling freely . we will show how to take these into account. each corresponding to a different value of . out the fact that this path is a parabola. We take as a trial we have to ﬁnd a function solution and quickly ﬁnd that ¤ d ) d ) r E ) C ( 7 ( ü d d ) ) ü ¥ d f " ) ! " g f E ) C ( E ) C ( U U is the initial height. we see that . we will simplify the problem by assuming that there is no friction or air resistance. We want a function whose derivative is and it is easy to see that (5.6). under gravity have the same slope The function which satisﬁes this differential equation is its solution and to solve the differential equation we must ﬁnd this function. In the case of a body falling under gravity. Some of the properties we could be interested in are the following: How long does the motion take? 3 E % C ( ( ) ¨ (5. There are many in straight lines with the same slope.66 FALLING BODIES In response to the second question we can try to deduce the physical laws which govern the motion of a freely falling body. Galileo around 1633–1638 was the ﬁrst person to gain a full understanding of the problem.1) with m/s.3. The reason that this constant appears is because the original differential equation (5.5) is another example of a differential equation. If we put equation (5. Equation (5. This illustrates the second of the two main uses of calculus. of the graph in Figure 5. He used experimental results to make general deductions. so that Galileo 1638 7 f 0 ¨ E ) C ð 0 (5. the velocity at time . Later in the book. He is supposed to have demonstrated this fact by dropping two cannonballs of different weights from the top of the Leaning Tower of Pisa and observing that they hit the ground at the same It has been observed that missiles and time. That is. Note that is a number.7) ü ¥ ! % ) ¤ ¤ 0 % ) ü d ) f ü E ) C ð ( E % C ü 0 ¤ ¤ E ) C ð ( 0 ms and Equation (5.
Thereafter. or is the independent variable.49 s. This is necessary. because Mathematica has no way of knowing which of the quantities .h[t]. but the expression sol[[1.1 did not take friction into account. As the body rises. due to friction slowing the motion.8) with we have and this is a simple quadratic equation with solutions s.8) gives the maximum m .GENERAL REMARKS 67 What is the maximum height reached? 3 In the case of the Tower of Terror problem.7) for the velocity gives Putting % E ) C ð ( % 7 % ) ) ! f ¥ U U ! U ! U ! ü ) % ! E ) C ! ð ( ) r " ( 7 ) ü ) ! ' " ¡ " ¡ 1 ( ) .1 The use of Mathematica Mathematica will solve differential equations and plot their solutions. so that we can make further use of it. m/s and ms into equation (5.8) with this value of the time to ﬁnd the height. the formula for the function. we need to ﬁnd the times when the body is 15 m high.3. at which time the instantaneous velocity is zero. the velocity becomes increasingly negative. Thus the total time to go s.1. . This is less than the actual time taken for Tower of Terror. because our up and down is hypothetical data in Table 5. which is about 6. The command % C ( U U ! ~ U U ) so that height: s. This seems physically reasonable. " We see that the time taken to reach the maximum height is half the time taken to rise and fall. we ﬁnd the time when the velocity is zero and then use equation (5.5 s. The free fall time is 4. Substituting this value of maximum height into equation (5. . that is.t] solves the differential equation and stores the solution as sol. its velocity decreases until the maximum height is reached. 5. To plot the graph we use the instruction E 1 . the initial height is 15 m and so to answer the ﬁrst question. The second question is also quite easy to answer. corresponding to the time that the body is back at 15 m. To ﬁnd the maximum height. ) g 7 E U ü d U f d U . r . The basic Mathematica instruction to solve differential equations is . We also have to take care to let Mathematica know that the function we want to ﬁnd is and that the independent variable is .2]] will extract . We also give the solution a name (sol). The ﬁrst time occurs when the body . corresponding to the starting point and s. ö f C % E with initial condition DSolve. Mathematica gives the following expression for sol: ð This is not a convenient form for plotting or further computation. This is a little shorter than for the Tower of Terror. is on the way up and the second time when it is on the way down. g d . Let us take the differential equation sol=DSolve[{h [t]==g*t+u. ) ) C ö U ( U f f C d " g r " ) ( ¨ ¨ . Notice that we need two equations in the instruction—the differential equation itself and the initial condition. From equation (5.h[0]==0}.
1. the height function is increasing over the ﬁrst part of the motion.6.0. 6. This produces the plot shown in Figure 5. EXERCISES 5. This is a particular instance of a general problem. 5. would you expect the time during which the ball rises to be longer or shorter than the time during which it falls? 3. It reaches a maximum value and then starts to decrease. Show that ( d ) ü d ) " g f E ) C ( is a solution of the initial value problem ( E % C ( 7 ü d ) f E ) C ð ( 2.{t.u>44}. How far apart are the stones 2 seconds after the second one has been dropped? 4.2]]/. Prove that the difference in the velocities of any two falling bodies is constant. Use Mathematica to solve the initial value problems: ' E % C { 7 " d ) { ¨ (a) ( How can you check the correctness of your answers? 5. A stone is dropped over a cliff. In Figure 5.10}]. If air resistance is taken into account. Let be a given function: % E % C ( 7 ' d ) g d § ) E ) C ð (b) U U d ) ! f E ) C ð ( Figure 5.{g>9.8. At the same time a second stone is thrown with a velocity which enables it to reach a height of twice that of the ﬁrst.4 INCREASING AND DECREASING FUNCTIONS Let’s look again at the motion of a body under the inﬂuence of gravity. Find the ratio of times for which the stones are in the air. a second stone is dropped over the cliff. A stone is thrown vertically upwards at 4 m/s.5. Two seconds later.68 FALLING BODIES 100 80 60 40 20 2 4 6 8 Plot[sol[[1.3 1. A ball is thrown vertically upwards.5: Mathematica plot of the solution to ) ¨ .
+ Y Find a number ! in the domain of for which This type of problem occurs more frequently than one might expect.6: The variation of height with time Find the portion of its domain where Y is increasing.6. Similarly. Let us clarify the precise meaning of increasing and decreasing functions. This has given rise to the wellknown greenhouse effect. The amount of carbon dioxide in the atmosphere is increasing as a result of industrial emissions. Á < 7 C . Y If we look at the graph in Figure 5. Y Y g Y g + ð Y + ð Y K  + Y M  Y The function we have + Y ! is said to be decreasing on if for any two points . achieves a maximum or minimum value. 2. The amount of carbon dioxide in the atmosphere is a function of time and is inﬂuenced by many factors. it is clear that the slope of the tangent is positive when is will be positive for all if is increasing increasing. we expect be proved and we shall assume them to be true. including legislation to reduce emissions. Current thinking is that the total amount of atmospheric carbon dioxide should be limited to no more than twice the preindustrial levels. They are summarised in the next theorem. which may have serious climatic consequences. Y Find the portion of its domain where Y ! is decreasing. Here are two instances: 1. Let the domain of a function .INCREASING AND DECREASING FUNCTIONS 69 Figure 5. and in with K  + Y K  The function we have + Y ! is said to be increasing on if for any two points . This suggests (but does not prove) that to be negative for all if is decreasing on . World population is expected to peak sometime around the middle of the 21st century and thereafter to decline. These facts can on . A good knowledge of the population curve is needed for planning purposes. it is necessary to be able to determine the effect factors such as restrictions on emission rates and reafforestation will have on the maximum value of carbon dioxide in the atmosphere. To ensure that this happens.
C < 7 Y ! be an interval in with . and in . .
E 2 C 2 2 2 2 2 2 d 2 ` f § 2 E 2 C Figure 5. If is a local extreme value of .2 Increasing and decreasing functions Let be differentiable on an interval . We begin with a deﬁnition. 2 E 2 C A local maximum or local minimum value of a function is often referred to as a local extreme value of the function.7: The graph of P U ' w 2 or " . . is . We shall assume that all our functions are deﬁned on an open interval or on the union of open intervals.70 FALLING BODIES THEOREM 5.4 Determine the intervals on which the function E 7 r C ' E 7 r % u E 2 C ð If 3 for all in . We ﬁnd that d 2 g " f 2 ' E 2 C ð 5. Consequently. then (a) increasing (b) decreasing. is is plotted in Figure 5. Solution.5 EXTREME VALUES Let us now turn to the problem of ﬁnding points where a function takes on extreme values. . The graph of 2 d E 2 ' ` f f 2 § C 2 E " f E 2 2 C C EXAMPLE 5. for all for ' u 2 u " % u E 2 C ð ' ' w g 2 u 2 " " u 2 ¨ " g ' U % w E 2 C ð ' From this we see that decreasing if u 2 u " if or and increasing if and if . It is said to have a local minimum at in its domain if all sufﬁciently close to . E 7 r C E 7 r C 2 % w E 2 C ð If 3 for all 2 in C . then E 7 r C is increasing on is decreasing on . the point is called a turning E E 2 C 7 2 C E 2 C E 2 C E 2 C A function is said to have a local maximum at in its domain if sufﬁciently close to .7.
8: Local maxima and minima point of the graph of .9. However.8. In contrast to local extreme values. where the function has and . If has a local extreme value at a point in its domain. if the curve is not smooth. The function shown in Figure 5. We state it without proof. then either or does not exist.8 suggests that for In Figure 5.3 Local extreme values Let be a function whose domain is an open interval or union of open intervals. A stationary point of a function is a point for which . Figure 5. other possibilities. does not have a global extreme value at either smooth curves the tangent is horizontal at a point where has a local extreme value. A stationary point which is neither a local maximum nor a local minimum is called a horizontal point of inﬂexion . Thus local maxima and minima of smooth functions are stationary points. in which case .EXTREME VALUES 71 Figure 5. + Y Y Y Y for all in the for all in  + 2 2  { + Y { Y ¨ ¨ Y .9: A function which is not differentiable at its local extrema or . but is not differentiable at these points. both graphs have a horizontal point of inﬂexion at .10. local extreme values at There is a theorem which formalises the above observations. THEOREM 5. + Y Figure 5.8 has a local maximum at and a local minimum at . This is illustrated in Figure 5. we can also consider point global or absolute maxima and minima. In Figure 5. + ð Y Y Y + ð Y + ð Y Y Y Y + ð Y  + Y Q  + Y A function is said to have a global maximum at in its domain if domain of . however. The word local is used here to indicate that we are only considering function values close to a in the domain of the function. There are. then there may be no derivative at a point where the function has a local maximum or minimum. It is said to have a global minimum at in its domain if the domain of .
which shows the graph of a function .4 The ﬁrst derivative test for maxima and minima Let be a function which has a stationary point at in its domain. so that when . 2 3 does not change as passes through has a horizontal point of will denote a If the sign of changes from negative to positive as local minimum value at .11. THEOREM 5. Since and . 2 2 If the sign of changes from positive to negative as local maximum value at . while on the interval . 2 passes through passes through . then . the function is by noticing that on the interval decreasing. 2 E 2 C ð 3 P 2 ¨ 2 P 2 2 E 2 C ð ¨ 3 . it curves down. we can make these terms precise . while number close to.72 FALLING BODIES Figure 5. then 2 EXAMPLE 5.10: Horizontal points of inﬂexion To determine whether a stationary point is a local maximum. but less than . 5.5. Consequently. has a local minimum at . On the interval roughly speaking. the graph. we examine the sign of the derivative on either side of the stationary point. the function is increasing. while on . We have . Similarly.5 Determine the nature of the stationary points of " d 2 ` f § 2 E 2 C Solution. 2 2 ¥ 2 2 2 E 2 C ð If the sign of inﬂexion at . we see that the sign of changes from negative to positive as passes through . The technical terms for this behaviour are concave up and concave down . curves up.1 Concavity E 7 r C % g u E f g C 2 ð g ® 2 % E 2 C g ð E 2 C ð ` f 2 ' E 2 C ð % w E Ú g C g ð . there is a local maximum at . a local minimum or a horizontal point of inﬂexion. Consider Figure 5. then has a has a . In the deﬁnition below. but larger than . This is the ﬁrst derivative test. (Why?) ð E ¤ 7 C E ¤ 7 ð C E 7 r C ¥ Ú 2 We shall use the notation to denote a number close to.
so that they are at right angles to the base. A beam with a rectangular cross section is to be cut from a circular log of radius cm. E 2 C { ¤ nc p a ve u P ¨ ve d nca o co co r wn decreases. cm.5 % u E 2 C ð ð E E C 7 C 2 g % 2 E 2 C ð ð % 2 w E ð 2 C % ð ð E 2 g C ð ð 2 ð ð In Exercises 1–4. Find the dimensions of the sides that will give the ' 2 % 7 " d 2 U f 2 E 2 C 9. 6. then up on . If the graph of is concave down on . Determine the dimensions of the gutter that will have the greatest carrying capacity. where the graph of has a point of inﬂexion at . 12. g g r " ( ( U g 10. 2 d " 8.11: Concavity on an interval Let be a function which is differentiable on an open interval . The does not imply that there is a point of inﬂexion at . In Exercises 5–8. An example of a point of result inﬂexion occurs in Figure 5. EXERCISES 5. A rectangle has a perimeter of largest area. Find the extreme values of the function 2 f " E 2 C  ` " f 2  E 2 C 7. The graph of is said to be concave down on if is decreasing on . The graph of is said to be concave up on if is increasing on . A point where the concavity of the graph of changes sign is called a point of inﬂexion . Suppose that is differentiable on an open interval . 2 ! " f ° 2 E 2 C ! f 2 U E 2 C 5. ﬁnd the intervals on which g d 2 ' f § 2 E 2 C increases and the intervals on which E 2 d " C § 2 E 2 C 1. An open gutter is to be made from sheet metal of width cm by folding up identical strips on each side.11. where is its breadth and its width. Note carefully that the converse of this result is not true. It follows that if the graph of is concave for all . . in which case . then for all . 4. 2.EXTREME VALUES 73 Figure 5. The . 11. ﬁnd the stationary points of and the local extreme values.  a d 2 g  E 2 C % ' f 2 d 2 E 2 C 3. Find the strength of such a beam is proportional to dimensions of the strongest beam that can be obtained.
.1. Height is the independent variable and is given in steps of 1000 m up to 10 000 m. where 1 mb is equivalent to 1 hPa. we denote the pressure by and the height by .1 THE AIR PRESSURE PROBLEM The measurement of air pressure is not a simple matter. Instead of bringing the time variation into our considerations. because air pressure at any given height varies with time. using these results to obtain a differential equation and then solving this equation to ﬁnd the required function that expresses height in terms of time. the height and the time . This same procedure will be used to deal with the second problem.CHAPTER 6 SERIES AND THE EXPONENTIAL FUNCTION In Chapter 5. but the average value is 1013 hectopascals. Technically. which is the pressure exerted by a cylinder of cross section 1 m and height sufﬁcient to give it a mass of 1 kg. By using this average we obtain the values given in Table 6. we shall look at the average pressure at each height. Indeed. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10 000 1013 899 795 701 616 540 472 411 356 307 264 Plotting against Table 6.1. this variation is one of the main factors determining weather conditions. Meteorologists usually express the pressure in millibars (mb). where one hectopascal is 100 pascals. We shall not be considering functions with two independent variables in this book and so we shall make some simplifying assumptions. air pressure is a function of two variables. for historical reasons air pressure is measured in hectopascals (hPa). the air pressure varies between 980 and 1030 hectopascals. which is about the maximum height of a commercial airliner. depending on the weather conditions. 6. In this table. The SI unit of pressure is the pascal (Pa). At sea level. but we shall require a new way of deﬁning functions in order to ﬁnd the solution.1: Average air pressure gives the curve shown in Figure 6. However. we solved the problem of bodies moving freely under the inﬂuence of gravity by considering some experimental results.
we get another curve rather than a straight line (Figure 6.1. so the approach of Chapter 5 fails in this case. If we do this. against and see if we get a straight line. but we can approximate it by measuring slopes of the graph in Figure 6.1: Variation of air pressure with height Here the problem is to ﬁnd the rule for calculating values of for given values of . the rate of change of pressure will also be high and. it will be more convenient to abuse the notation and take the approach we did in Chapter 5 (page 58). so we might try the same approach as we did in Chapter 5.2: The derivative of the pressure If we plot the graph from this table. as described on page 61. It also suggests that the rate of change of pressure depends basically on the pressure rather than the height. we obtain the following data. It may denote either a We will write number or the rule for calculating this number. We cannot compute from the given We plot data. . What do we try next? Let’s think about it. a given volume of air weighs less and so its contribution to the overall pressure is less important.2). The symbol is thus doing double duty. The pressure depends on the mass of the column of air above a given point and as we go higher the air gets thinner because air is compressible. the rate of change should also decrease. We group the plot points in threes and take the slope of the chord between the ﬁrst and the third point as an approximation to the slope of the tangent at the middle point. The function is expected to be smooth. 0 f g " % f ( 1 ¨ ¨ ( 1000 f " " % 2000 f % % 3000 f ! % % 4000 f % ! % % 5000 f " a % % 6000 f U ` % % 7000 f ! r % % 8000 f g r % % 9000 10 000 g U % % f ` U % % Table 6.76 SERIES AND THE EXPONENTIAL FUNCTION Figure 6. so that we need a function which enables us to express the pressure in the form . This means that as we go higher. This suggests that when the pressure is high. as the pressure decreases. instead of . We have done this by taking central differences . In what follows.
( ð % 1 % E % ( % C " 1 % % E % 1 ! ( C 1 1 % % % ` 1 % % % U E ( C ( % % % g E ( ( ¨ C ~ 1 1 ¨ 2 % % % % % r " .
ð 5 . we try plotting against and see what happens. We don’t know in advance if this will achieve anything.THE AIR PRESSURE PROBLEM 77 Figure 6. but it illustrates the way in which we must use all our physical intuition about a given situation in trying to produce a workable mathematical model of it.2: Variation of the derivative of the air pressure with height Given this argument. In this case we do indeed get a straight line (Figure 6.3) to within a small amount of error.
Rather than attack equation (6. the constants are less friendly than the previous cases. It is more complicated than those we encountered in Chapter 5 for two reasons. gives us the initial value problem + % B + B + ð It is this initial value problem that has to be solved for the pressure as a function of height. together with the initial value .1) directly. This seemingly minor matter actually complicates the solution considerably. First. 3 3 2 Figure 6.3: Variation of the derivative of the pressure with pressure + The equation of this line can be obtained and this equation. the function as well as its derivative appears in the equation. More importantly. This is often a useful procedure in . we consider a simpler form.
2 6 5 5 6 5 6 6 5 6 5 < < 6 5 6 6 5 7 7 6 6 5 6 5 C C 5 5 H H H H 5 .
3 3 B B B B 2 (6.1) .
In words. We take a general approach and put and. we 2 ¤ d ¤ d 2 2 d g r § d 2 r 2 2 r ¨ ' ¤ d 2 2 ¨ g d 2 r ' . Although this will not work here for reasons which will soon become clear.2) { 1 E 2 C { " E % C { 7 { { 2 ¨ ¨ To solve this problem we need to ﬁnd as a function of . This is a typical instance of the way mathematical and scientiﬁc ideas develop in tandem. the left hand side becomes 2 { ¨ r { d 2 r { while the right hand side is We then have and so equating coefﬁcients gives and . Equation (6. a similar problem arises. Thus for all values of .2). we . and .2) becomes try % E 2 C { r % r { ¨ ¨ d d 2 ¤ 2 r d 2 r 2 d § 2 " r E { % C { d 2 r { while the right hand side is We then have (6. Examine an easier version of the problem ﬁrst in order to gain the necessary insight before attempting the more difﬁcult problem. but does not satisfy the initial condition . There is more than one way of doing this and we shall consider two methods in this book. if is equal to its own derivative. where and are constants. where we solved differential equations by using polynomial functions as trial solutions. If we substitute this into the ﬁrst Take a trial solution equation of (6. it is still useful to see what happens if we attempt to use polynomials as trial solutions. We ﬁnd out why they don’t work and what we have to do to actually get the solution we are seeking. The methods developed will prove to be useful in later problems.78 SERIES AND THE EXPONENTIAL FUNCTION difﬁcult problems. This is a solution of the differential equation. namely the problem of ﬁnding from this equation.3) ¨ d 2 ¤ d 2 d § 2 r { Once again. . That is. We shall consider the special case consider the initial value problem " E % C 1 7 E ( C 1 E ( C ð 1 ¤ r . and are constants. If we equate coefﬁcients " ¤ " % ¨ d r where . then (6. this problem is to ﬁnd a function which is equal to its own derivative. If we try a higher degree polynomial.1) has the form 7 ¤ E % C 1 7 E ( C 1 d r E ( C ð 1 We now have a problem which is purely mathematical. A physical problem gives rise to a mathematical problem and subsequent investigation of the mathematical problem may shed new light on the physical problem. for example. there is a discrepancy between the degrees of the terms on the two sides of the equation— the left hand side has degree 2. We begin by trying the approach we developed in Chapter 5. one in this chapter and another in Chapter 10. Then the left hand side of the ﬁrst equation in (6. Suppose. while the right hand side has degree 3.
If . it quickly follows that and we cannot satisfy the initial conditions.5). so that Ê d ¯ ¤ % 2 d ¯ ¤ 2 " Finally. If we substitute equation (6. but for the moment let us press on with trying to ﬁnd a solution to the initial value problem (6. ¯ r ¦ ¦ ¦ ¥ ¤ r ¤ r . we evaluate the constant by using the fact that in each side of equation (6. We will ﬁnd that .4) is an inﬁnite sum.7) . . This problem of running out of terms on one side of the equation will occur whenever we take a trial solution of the form .6) (6. that is. Equating coefﬁcients in this equation gives us r r . This is an assumption which we shall have to examine later. The right hand side of equation (6.THE AIR PRESSURE PROBLEM 79 (Theorem 5.4) term by term to get equation (6.1). suppose we try a solution of the form (6.5) (6.3) leading to the result . we get E constant terms 2 C ¯ r § There is a promising pattern emerging which is unfortunately spoiled by the fact that there are not enough terms on the left hand side of equation (6. suppose we use a trial solution which never runs out of terms. Such an inﬁnite sum is called an inﬁnite series and we shall soon have to be quite precise about the meaning and deﬁnition of an inﬁnite series.2) we ﬁnd d ¤ 2 Ú ¤ r E " d C d ¥ ¤ 2 ¤ r d ¦ ¦ ¦ d 2 § r ' d 2 r g d r 7 7 g 7 % " 7 % r ± % % r 7 r d ¤ 2 ¤ r d ¦ ¦ ¦ d ¤ 2 2 ¤ r r d d 2 ¥ ¤ r 2 d ¥ ¤ r r d ¦ E ¦ 2 ¦ C % { d ¨ 2 r d ¤ 2 r d r E 2 C { We have assumed that we can differentiate equation (6. Putting d ¤ 2 E ¤ § r 2 coefﬁcient of C E 2 coefﬁcient of C E d 2 ¦ ¦ ¦ d ¦ d d ¦ ¦ ¦ g ¦ g ¦ r r 2 ' d d ¯ ' 2 d ' g g r r d 2 ' d r g % ¨ 2 g g ¨ g ¨ r d ' ' d r d ¤ r r 2 ¤ g 2 g d § ¤ d " ' r " É r ' E 2 C E { r 2 C { (coefﬁcient of (6.6) gives .2).4) into equation (6. However.4) where the terms go on forever.2)). leading to { § 2 " 2 r r % 2 when (equation (6. . rather than a ﬁnite sum. . . .
1 1. the terms all appear to approach a ﬁxed value as becomes progressively larger.2 2. for a given . The following table gives the density of the atmosphere at a height . we will consider inﬁnite series in more detail.80 SERIES AND THE EXPONENTIAL FUNCTION At this stage it is not clear that we can get any useful information out of equation (6. so that for a ﬁxed value of . The calculations we have carried out here are all exploratory and suggestive of methods which . Notice that in equation (6.38 20188. determine a differential equation relating the height and the density.68 0. Based on this data. In the case of equation (6. We can illustrate (but not prove ) this in the following table.56 0.12 0.3: Partial sums for various values of and % % " % r r " ¤ % 2 for three values of and various increasing values of . What we hope is that they become so small that we can ignore all terms after a certain number and still attain the accuracy we want.0078 36. because the additions go on forever.5 2. deﬁned as an inﬁnite sum. The density is given as a fraction of the mean sea level density.125 15.333 2. we would expect that eventually successive terms will become smaller and smaller.84 0.6 2.7). for different values of . the denominators of the terms become larger as we move to the right. Our task now is might be used to ﬁnd a solution of the initial value problem to investigate how to make them precise.28 2 ¯ 2 d ¯ Table 6.71828 148. How can we operate sensibly with an inﬁnite number of terms? We certainly cannot add up an inﬁnite number of terms.71828 143.70823 65.0039 42.96 0. while the height is given in kilometres. with a view to justifying these assertions. In part.5 ( ( 0. the answer lies in the fact that in any practical calculation we only want accuracy to a certain number of decimal places. In the next section. but it does suggest that if we ﬁx a value of .52 5.0019 47.7).24 0.5 1 % " r 1 Notice how. A table such as this is not conclusive. then by taking enough terms of the series (6.0312 26.413 22026.413 22026.689 10086.0625 21. We have used Mathematica to calculate the ﬁnite sum 2 ¤ 2 d ¦ ¦ ¦ d § 2 % ' " d ¦ d ¦ ¦ g 2 d 2 g 2 r d d 2 " d " E " E 2 C 2 C { 1 " 2 6 11 2.0156 31. 0. The technical term for this phenomenon is known as convergence. We have to answer questions such as: How do we add up an inﬁnite sum? 3 " E % C { E 2 C { E 2 2 C ð { { 2 Can we differentiate an inﬁnite sum of functions in the same way as a ﬁnite sum? 3 Is there a function. which is a solution of the differential equation? 3 EXERCISES 6.71828 148. ¯ ¯ 2 2 2 ( .25 10.375 644. .40 0.7) we can calculate to any accuracy we want.7) this is indeed what happens.71828 148.
It should be noted that in everyday usage. We Ñ W r d ¦ ¦ ¦ d r d d " " " g g g g f d d d " " " " 3 3 3 3 . There are two types of notation for series.2 INFINITE SERIES We begin with a provisional deﬁnition: Provisional deﬁnition 6. This is the way we have written series in the r discussion above. sequences are almost always referred to as series. while ¤ is used to denote the inﬁnite sum ¤ ÷ W is used to denote the . The ﬁrst is to write down the ﬁrst few terms of the series and use dots to indicate that the sum does not terminate. The ﬁrst is the sequence of terms.INFINITE SERIES 81 6. To say what is meant by the sum of a series requires the theory of sequences which we discussed in Chapter 2. The numbers in the list are called the terms of the sequence. The symbol ﬁnite sum r EXAMPLE 6. There are two sequences associated with every series.1 Some particular series are: " or g ÷ Ñ ¦ It is somewhat misleading to say that a series is an inﬁnite sum. The numbers which appear in the sum are called the terms of the series.1 A series is an inﬁnite sum of numbers. a list of stock market prices is referred to as a series of prices. We recall the following facts on sequences: A sequence is an inﬁnite list of numbers. The second is to use summation notation. because there is no way to add up an inﬁnite number of terms in the same way that we can do for a ﬁnite sum. For example. ¤ r Ñ Ï Î Ð Í ¤ ¡ ¤ r 3 A sequence be divergent. 3 is said to be convergent if exists. in the series 7 ¦ ¦ ¦ d " ! d " U d " g d " E ¥ g C ~ " Ñ W i E ± " f C or ÷ ¦ Ú " or ± ÷ Ñ ¦ ± or Ñ ÷ ¦ ¥ ± ÷ Ñ W W W W ÷ ¦ ¦ ¦ ¦ ¦ ¦ ¦ ¦ d d d d " " " ! U U U f d d d " " " ' ' U ' d d d r may replace by a formula involving in both the above cases. For example. ¦ ¦ ¦ d § r d r d r r . Otherwise the sequence is said to It is essential to be clear about the difference between sequences (lists) and series (sums).
we look at the sequence of partial sums Õ ¤ 2 f ¤ 9 There are two cases to consider: l " ¤ 9 l 2 " u 2 u " f " u  2  If  3 .82 SERIES AND THE EXPONENTIAL FUNCTION the sequence of terms is ¦ ¦ ¦ 7 " ! 7 " d 2 f " 7 " " d " % 2 d Ô U 9 2 7 7 d r 2 " g 7 d " ¤ 9 ¤ ¡ " ÷ d " The sequence of terms is important in the theory of series. . then diverges and so the sequence of partial sums also diverges.1 Partial sums: The th partial sum of the series and so by ¦ ¦ ¦ d r d r ¤ 9 ¤ 9 is the sum of the ﬁrst terms of the series. so that " for the sum of the series (6. then as .2 Consider the series ¦ ¦ ¦ d § 2 d 2 ¥ 2 Ñ W ¤ 9 Ñ Ð ¤ Ï Î Í Ñ X ¡ ¤ 9 ¡ ¤ 9 r ÷ X The th partial sum of this series is ¥ ¤ ¥ ¤ W This is a ﬁnite geometric series with sum ¤ 2 9 To determine the convergence properties of the inﬁnite series. 7 ° 9 7 § 9 The partial sums of a series form a sequence us to say what we mean by the sum of a series.8).8) diverges if . but much more important is the second of the two sequences associated with a series. It is denoted (6. except in some special cases.2 Sum of a series: Let be a series and let be its sequence of partial sums. A series which is not convergent is divergent. Ñ . If the series is convergent. " 2 ¡ ¤ 2 "  2  3 2 f " ¦ ¦ ¦ d § 2 d " " u  2 if we can write 2 f " 2 d ¦ ¦ ¦ d § 2 d 2 2 f 2 f r ¤ ÷ r 2 d X ÷ DEFINITION 6. It is not usually possible to ﬁnd a formula for the sum of a convergent series. that is. If . EXAMPLE 6. This means that the series (6. Ñ . and it is this sequence which enables DEFINITION 6. This is the sequence of partial sums.8) ÷ ¤ 9 as . then is called the sum of the series and is denoted by . The series is said to be convergent if the sequence is convergent. Thus .
f e d c r " " d ¦ ¦ ¦ d " a d " r d " 15. ' ÷ Ñ W ¤ 8. then our conclusions in Section 6. some series cannot be expressed in terms of these known functions and it is in such cases that our main interest lies. The series ¡ " h d ¤ 2 d ¦ ¦ ¦ d § 2 d 2 d 2 d " E 2 C { E E " 2 7 " f f " C C ~ " which occurred in our discussion of atmospheric pressure is just such a series. ÷ ¤ E i ¦ ¯ ¦ ¦ d § 2 § ° ÷ ÷ d 2 W W f ¯ 2 ' " d 2 d g " E 2 " C 7 " Thus if is the function deﬁned by interval and for any f C g 2 E " 7 " f C . " Ú ¤ d ` g § ÷ Ñ W Y ) g ) E " f C 5. Ñ ÷ ) " 6. then the domain of is the open " d ± g § § § ÷ ÷ ÷ W W W W . However. ' d " d ¦ d ¦ ¦ a d d " r d d " ' d d " " d r d U d ' d g d " 13.1 will have to be reexamined. although when the series form is used for the deﬁnition. It is just the simple rational function . ¤ " U ¤ E " f C ÷ ¤ ¤ " 11. " E 2 C E In the above example the function we have produced is not new.2 Evaluate the following sums. Ê ¤ ' g É ÷ Ñ W ¤ ¥ ¤ " 7. 16.INFINITE SERIES 83 . E " f ± g C 2. 14. C ÷ terms ¤ ³ 2 a d ° 2 ` f § 2 r d 2 U f 2 ' d % " g d g " " d " ` d " 17. We shall come to the reason for this later. g 18. ¯ " ³ ÷ ÷ W ¤ Y ' 3. ÷ ) 4. If the series does not converge. EXERCISES 6. This example also shows that a series can be an alternative form of a function we already know. U ° ÷ ¤ 12. ÷ Ñ W 10. but this is something we have to verify. Ê ¤ g ' f ¤ r g É ÷ Ñ W ¤ " f ¥ ` ' 9. the domain is reduced from to . In that discussion we implicitly assumed that the series converged. 1. b Express the following sums in notation.
 d r + ¤ ÷ W r 19. ¤ W 7 B ÷ ¤ X r ÷ ¤ X Suppose that ¤ W and . so that indirect methods are necessary. ¤ ÷ 20.3 CONVERGENCE OF SERIES There are a variety of methods for determining whether a given series converges or not. ÷ 22. Show that d C + d ¦ ¦ ¦ d d C d 24.2 we computed the sum of a series by computing a sequence of partial sums and then ﬁnding the limit of this sequence of partial sums. Verify that d B d + Hence show that d + Ñ W 26. In Example 6. W . Such direct methods are usually not possible except in a few simple cases. Verify that d C B B C d C + < B C + Hence evaluate B < 7 Ñ W 6. Sum the series Ê C d B d É Ñ W 25. Find the values of the following sums. Verify that C d B d C d + d + Hence evaluate C d + d + Ñ W 27.84 SERIES AND THE EXPONENTIAL FUNCTION 23. One of the most useful and important tests for convergence is the ratio test which we state without proof. H r ÷ ÷ ¤ ÷ ÷ ¤ ¤ ÷ ¤  r B C + 21.
we can compute to any required accuracy. Suppose that the sequence Õ Ú ü ° ü § ü ü Let us apply this test to the function deﬁned by (6. For any . if we take enough terms of the series. it follows that the series converges for every and so the domain of is .CONVERGENCE OF SERIES 85 THEOREM 6. If " w 2.9) may converge or it may diverge. We want to determine so that E 2 C 2 ¤ 2 § 2 2 2 h g 2 h gives the sum of the series ¤ 2 § 2 2 to within the required accuracy.10) § 2 2 Since this holds for every value of . If . The series (6. This raises an immediate question. How do we determine when enough terms have been taken? Let us put this another way.9) converges. . then no conclusion can be drawn. Let us write ¤ 2 § 2 2 ¦ ¦ ¦ d " d ± ú ú ú ¯ ¯ ± 2 ¯ Ñ ú ú ú Ï ¦ Î Ð Í ¦ Ñ d ¦ d  Ï d ¦ ¦ Î Ð Í ¦ ¦ 2 ¦ d  % ¦ ¦ ¦ ¯ ¯ ¦ E E d d " " ú ú ú ú d ¯ Ú ¯ ¯ ¯ d E ¤ f ' ' " ' 2 ± C C d d ¯ f d ~ ± ± d g ~ g C ¥ g ¯ ~ ¤ ¤ 2 d 2 d 2 For a particular value of . the th term is ± 2 and using the ratio test we have ¦ ¦ ¦ d ¯ ' d ¯ g d 2 d " d ¥ 2 2 d d 2 2 d d ú ú ú ú ¤ ¤ E d " 9 9 " 2 " Ñ C Ï Î Ð Í E E 2 2 C C " 3. ú ú ú ü ú ú ú Ô ¦ ¦ ¦ 7 ú ú ú § ü ú ú ú 7 ú ú ú ü ú ú ú 7 ÷ ú ú ú ü ú ú ú . then the series (6. . If " u 1.9) be a given series of nonzero terms.9) diverges. ú ú ú ú ú ú ú ú converges to a limit .1 The ratio test Let ¦ ¦ ¦ d § ü d ü d ü ü Ñ W (6. then the series (6.
86 SERIES AND THE EXPONENTIAL FUNCTION We can’t calculate this exactly. it is a bit easier to compute the error. we ﬁnd . So if result smaller than the allowable error . it can be shown that the error made by taking a given number of terms is always . The series ¯ ¤ ( Ñ 9 ¦ ¦ ¦ d ¯ ¤ ( d ¦ ¦ ¦ d ¯ § ' ( d g ( d ( d " E ¡ (6. but we can estimate it and as long as we take care to overestimate it. The signs of the terms in the series alternate and when this occurs for a convergent series in which the terms are decreasing in absolute value.11) Ê E g ¦ Ê ¦ d ¦ ¦ ¦ ¦ C ¤ d E 9 ¯ " d E " § § d Ú 2 E d ¤ C 2 2 d C C This shows that the error made in approximating ¤ 2 by is E 2 d " 2 ¯ d d f 2 ~ C ¤ ¯ 2 d " E 2 ¤ " " f É " f ¯ ¯ ¯ ¤ ¤ C 2 2 C .11) with ¦ ¦ ¦ d ¯ ¤ 2 d ¦ ¦ ¦ d ¯ § ' 2 d g 2 d 2 d " E E C 2 EXAMPLE 6. Then ¯ ~ ¤ 2 ¥ % ¤ 2 u 2 d ¦ ¦ ¦ d § 2 d 2 d 2 ¡ d " E 2 C % u 2 with an error of less than . (6. ¡ ÷ ¯ E ¤ g " a ! f U ` C " ¯ a " ´ g ¯ d ' ¦ ¦ ¦ d g " g d " Ê " g É ( C { a Thus we use and compute a ` if if g 2 ' % % % % % % % % % ¦ ¦ % % v u E g ~ " f ¤ g C ~ ¯ " " E % " % % f % ¦ C % ¤ ¡ with an error of less than Using inequality (6.12) i with an error of less than . to within the required degree of accuracy. For positive we can be sure that we have computed values of we have ¦ ¦ ¦ d 2 d 2 E 2 C d " É ¤ 2 ¤ 2 If in equation (6. At this stage you should reexamine Example 2.3 Suppose we want to compute for the function u  2  if 7 .10).16 of Chapter 2 in the light of the above discussion. compute until we get a smaller (in absolute value) than the ﬁrst omitted term.
We now reach a very important deﬁnition. we would like to know how to differentiate any series. This is indeed true in this case. We have to show that E 2 C ð { E % C O Ó N M h E 2 C O N M " Ó E % C O N M 7 E 2 C O N M E 2 C E O 2 N C M O N ¨ M 2 ! g ! " a g Ó " E % C E { 2 C E O 2 C N M { The ﬁrst of these equations is the one that will cause concern. More generally. We were led to the exponential function because of our interest in the initial value problem . We conclude this section with some additional examples on the use of the ratio test. where is a constant The domain of the exponential function is and by calculating whose value is for various values of .5).4) to equation (6. as in Figure 6. we assumed that we could differentiate the series term by term.CONVERGENCE OF SERIES 87 Figure 6. in principle. we often write for . .4: The exponential function has been shown to converge for each value of .3 The exponential function: The exponential function is deﬁned by ¤ 2 Ñ 9 ¤ 2 § 2 2 E 2 C O N M ( For reasons which will become clear later on. This will be discussed in the next section. assuming that such a series is differentiable. We need to be precise about what we can and cannot do when differentiating series. a simple matter. To show that satisﬁes this initial value problem is. we can plot its graph. In going from equation (6. ¯ P L J ÷ I ¤ g G ¦ E ¦ ¦ D d ¯ " d ¦ ¦ ¦ d ¯ ' 2 ¨ d % g ¨ " r g ` ' U d 2 2 d ¨ " " E f 2 C O N M g f . DEFINITION 6. The value of is easily seen to be 1. In order to evaluate 7 E 2 C O N M ¨ we have to know how to differentiate the series which deﬁnes the exponential function.4. but is not true for series in general.
5 In the series ¤ C Ñ 9 we have Ú ¤ d C + Ú ¤ ü % ¤ C ¤ ü Hence § Ú ¤ C Ï Î Í Ú ¤ ü Ï Î Í Since this is greater than 1. We have ¯ d + Ú ¤ ü % ¯ ¤ ü Hence ¯ § Ï Î Í Ú ¤ ü Ï Î Í Since this is less than 1. the series will converge.4 Determine whether or not the series Ñ 9 converges. EXAMPLE 6. EXAMPLE 6.6 In the series Ñ 9 we have d C Ú ¤ ü % B C ¤ ü ú ú ú ú ú ú ú ú ¯ ú ú ú ú ¤ C § ú ú ú ú ¯ ú ú ú ú ¯      d d d d d C + + + + + ú ú ú ú ú ú ú ú ¯ ú ú ú ú ú ú ú ú ú ú ú ú B Ñ Ñ Ñ Ñ Ñ Ï Ï Ï C Î Ð Ð Î Î ÷ Ð Ð Ð Í ÷ Í Í C ¤ ¤ ¤ ¤ ¤ ¤ ¤ ÷ ¤ ú ú ú ú ú ú ú ú ¤ ¤ ü ü ú ú ú ú ú ú ú ú Ñ Ñ Ð Ð ¤ ¤ .88 SERIES AND THE EXPONENTIAL FUNCTION EXAMPLE 6. the series will diverge.
the Swiss mathematician Leonhard Euler proved the following result. calculate the value of the partial sum places and use the ratio test to decide whether or not the series converges. The number is normally associated with circles and the numbers arise in conjunction with squares and yet we somehow have a connection between all of these numbers using an inﬁnite series. " " This is a bizarre and totally unexpected result. How accurate is the approximation? U 7 " ¤ r d U d f g C ¦ ¯ ¦ r ¦ ' ¦ " 9. 9 ¤ ¤ ¯ E ¯ 3. 9 9 f d g g g g g ú ú ú ú ú ú ú ú ú ú ú ú Ñ Ñ Ñ Ï Ï Î Î Ð Ð Ð Í Í g g " ¤ ¤ ¤ ¤ ¤ ' Ñ 9 9 9 9 9 correct to four decimal . ÷ Ñ ¤ 6. g r d g 1. ÷ Ñ ¤ 2. ÷ ¤ 10. Show that f ¯ g 2 d 2 f 12.CONVERGENCE OF SERIES 89 Hence " ¯ E " § 9 d ú ú ú ú f C " ` " ¤ Ú ¦ ¤ E ¯ d ¤ ¦ " E Ê E g E ¦ ¯ ¤ " g ¦ " ¤ ú ú ú ú % g C § ¦ d g f ú ú ú ú C É C " ¦ C ¯ " " " § ~ ~ ' d 2 " " ÷ ÷ ÷ ÷ ÷ Ñ Ñ Ñ Ñ Ñ f d d " Ï Î Í Ú ¤ ü " ü Ï " ú ú ú ú " ' d ¤ ¥ g ú ú ú ú Ó 7 d Ñ E " Î 7 Ð " Í Since this is equal to 1. EXERCISES 6.3 For each of the following series. 9 ¤ ¤ ¯ % 7. In 1734. ¤ 11. g C ÷ Ñ ¤ 4. we can draw no conclusion from the ratio test. " ÷ Ñ ¤ 8. 9 ¤ r d r ¤ U Ú ¤ E " f C 5. Use the ﬁrst 10 terms to approximate .
to approximate . 2 Ú ¤ r Ï Î Í Ú ¤ ¤ 2 Ú ¤ r Ï Î Í Suppose that there is a positive number such that Ú ¤ r Ï Î Í 3 and so the series will converge by the ratio test. deﬁned by a power series: § § Ó are used to approximate approximation? . We write . Suppose that the ﬁrst 6 terms of the series ¤ 2 § 2 2 6. The interval is called the interval of convergence. The series converges for all and diverges for all such that . which are of the form 7 ¦ ¦ ¦ d 2 r d ¦ ¦ ¦ d § 2 r d 2 r d 2 r d r 2 r Ñ W To determine where this series converges we apply the ratio test. that is. E l 7 l f C ú ú 2 ú ú l 6 7 " u  2  ú ú ú ú ú Ú ¤ ¤ ú ú ú 2 2 ¤ Ú r ¤ r ú ú Ñ Ï Î Ð Í ¤ ~ " u  2  2 If we choose so that . Consider a function 7 ' 7 g 7 " ± r are all constants. assuming that all the coefﬁcients are nonzero. In particular. The interval of convergence is . Calculate the d ¯ ' d ¯ g d 2 d " ÷ . Use the ﬁrst 5 terms of the series expansion of maximum possible error in the approximation.4 RADIUS OF CONVERGENCE From now on we shall focus on series of functions. we shall consider power series. then ú ú ú ú There are three cases to distinguish. What is the value of the approximation? How accurate is the E U % C y x w ¦ ¦ ¦ d ¯ E 2 C d ¦ y ¦ x ¦ w 13. ¦ ¦ ¦ d ú ú 2 ú ú ú ú r ¤ Ú ¤ r d ¤ r ¦ ú ú r ¦ ¦ Ñ ú ú ú ú Ï Î Ð d ú ú Í Ñ ¤ §  Ð 2 ¤ 2  § ¤ r r ú ú d ú ú ú ú Ñ 2 Ð ¤ r 2 ¤ d r 2 ú ú ú ú r Ñ d Ð r ¤ E 2 C where . rather than a number. The number is called the radius of convergence and is usually denoted by the symbol . each term in the series is a function. If 7 % ú ¤ Ú ¤ r r ú Ñ Ï Î Ð Í ¤ 3 6 u  2  ~ " E 6 7 6 f ú ú C 6 ú ú 6 w  2  2 then the series converges for every value of and we say that the radius of convergence is inﬁnite.90 SERIES AND THE EXPONENTIAL FUNCTION 14.
We shall not investigate this issue in any detail. at the points . case some of the constants This does not cause any problem and we shall deal with such cases as they arise. However. There ú ú Ï Î Ð Í ¤ . Such series still have a radius of convergence and there will usually be no difﬁculty in determining it. There are three further points we should note: The series may or may not converge at the end points of the interval. . The series will still have a radius of convergence. There is no problem with any and we might have expected the series to converge for all negative values of . since the interval of convergence of a power series has to be a symmetric interval about . but it will be more difﬁcult to determine. so the interval must be . We have Ú ¤ 2  2  Ñ  Ï ú ú Î 7 Ð Í 2  " ¤ u ú ú ú ú ú ú ¤ 2 % ú ú ú ú Ñ 2 then the series converges only for is no interval of convergence.8 Determine the radius of convergence of the series 2 ° 2 2 ¤ 2 Ñ W E " 7 " f C " 2 " % u 2 2 2 2 d " f § g " d ¦ ¦ g ¦ d d § g 2 d d " 2 d 2 ¤ g E d " " 7 ÷ " ¤ f C as long as is in the interval We also know that 2 .7 We have already discussed the series ¦ ¦ ¦ d § 2 d 2 d 2 d " Since we cannot put in the right hand side. the series cannot converge for this point. In this will be zero and the above arguments will need to be modiﬁed. We shall meet cases of power series where only even or only odd powers of occur. We say that the radius of convergence is zero. we conclude that every power series has a radius of convergence and the series converges in a symmetric interval about the origin. that is. It may be that does not exist. 2 r  ¤ r ~ Ú ¤ r  Ñ Ð ¤ Ï Î Í 6 ® 2 EXAMPLE 6.RADIUS OF CONVERGENCE 91 If 7 l ú ¤ Ú ¤ r r ú Ñ Ï Î Ð Í ¤ 3 3 3 3 From this. Let us . and the series diverges at EXAMPLE 6. the origin. This will not occur in any of our examples.2 and the results we got there show that its interval of convergence is derive the same result by applying the ratio test to the series. E " 7 " f C in Example 6.
Hence the radius of convergence is ú ú ú ú ¤ ¤ g 2 § U Ú Ú ¤ ú ú ú ú ¤ T g g 2 2 ú ú ú ú ú ú ú ú Ñ Ñ Ï Ï Î Î 7 Ð Ð Í Í " ¤ ¤ u ú ú ú ú ¤ Ú ¤ ü ü ú ú ú ú Ñ Ï Î Ð Í ¤ 9 9 9 9 9 9 9 9 . ¯ ÷ ¤ 10. ¤ 2 ¯ ÷ Ñ 9 ¤ Ú ¤ E " ¤ d g C 2 7. ÷ Ñ ¤ 14. ÷ Ñ ¤ 6. Ñ ¤ Ñ Find the interval of convergence of each of the series in Exercises 11–15. ¤ § 2 E g f ' C ¯ ¦ E ' C ¦ a ¦ U ¦ " ÷ 9 ¤ ÿ g 2 þ ¯ E " E d " f C C 9.4 Find the radius of convergence and interval of convergence of each of the following power series. that is. ÷ Ñ ¤ 2. Ignore the problem of convergence at the end points.92 Here SERIES AND THE EXPONENTIAL FUNCTION EXERCISES 6. ÷ Ñ ¤ 4. ¤ E g d 2 C ¤ r ' ÷ Ñ 9 ¤ ¤ E " f ¤ 2 C E " f C 11. Ú ¤ Ú ¤ 2 U ¤ E " f C ÷ Ñ 9 ¤ ¤ 2 ¤ g 3. ¤ ' ÷ Ñ ¤ 8. if . ¯ E " d 2 g C ¤ E " f C ÷ Ñ 9 ¤ Ú ¤ ¤ 2 r 5. ¤ ¤ 2 § ¤ ' ¤ E " f C ÷ Ñ 9 ¤ ¤ 2 ¯ g 1. Ñ ¤ Ñ ¤ E ' f 2 C ' E " ¯ f g C 15. ÷ Ñ ¤ 12. ÷ ¤ ¤ E " d 2 C " d ' ¤ E " f C ÷ ¤ ¤ ¯ E g ¤ d g 2 C 13. . Ignore the problem of convergence at the end points. g 9 g u 2 u g f g u  2  as long as .
an inﬁnite sum of differentiable functions may not even be continuous and therefore certainly not differentiable. and so is not differentiable there.2. We know from Theorem 4.2 The derivative of a power series Let be a function deﬁned by a power series ¦ ¦ ¦ d 2 r d ¦ ¦ ¦ d r E 2 C This result is not surprising. It tells us how to differentiate a function deﬁned by a power series inside its radius of convergence. Then the function E § " E r ± d 2 ¤ 2 d " is differentiable everywhere in the . not a power series. A simple example is the series 2 This is a geometric series and for any .] 6. However. of course. it is not true for all inﬁnite sums of functions. It has obvious practical implications.2 that the derivative of a ﬁnite sum of functions is the sum of the derivatives and the result above tells us that the same is true for an inﬁnite sum of positive integer power functions. using the result of question 7 in Exercises 6. The result is as follows. show that Ñ 9 We could compute to a given accuracy by using either one side or the other of the equation above. It is. Hence the function is not continuous at . but the proof is rather difﬁcult and we shall not present it here.5 DIFFERENTIATION OF POWER SERIES We now present a very important result. ¦ ¦ ¦ d ¥ 2 and let the radius of convergence be interval and d r E 2 C ð E 6 7 6 f C . [This is a technique known as improving the convergence of a series. Which expression would be more efﬁcient for doing this? Explain your reasons. Verify that E " d " C d E " d " " d 2 C d d ¦ ¦ § " ¦ C 2 § d " d r ÷ Ñ 9 ¤ d 2 E § ÷ Ñ d 9 ¤ r 2 2 " ' % 2 d r C d % " k d C 2 " w 2 2 r d g 6 r ÷ " Hence.DIFFERENTIATION OF POWER SERIES 93 16. its sum is 1. % 2 % 2 ¦ ¦ ¦ d 6 The radius of convergence of the series for the derivative is also . then the sum is 0. The result is proved using differentiation by ﬁrst principles. In fact. THEOREM 6. but it does tell us something very new. If .
3 The derivative of the exponential function .2 can be written in the form starts from 0 in the ﬁrst E 2 C { E 2 C ð { We can now see that the exponential function is a solution of the initial value problem . we have For all E 2 C y x w 2 ¨ ¨ h g 2 h g ± 2 y x w ± of Section 6. but we are making progress. Let 7 2 r Ñ W ¦ ¦ ¦ d 2 r d 2 r d r E 2 C ± for .2. . 2. The only functions which have this property are the functions . EXERCISES 6. THEOREM 6.1. explain why the index of summation sum and from 1 in the second sum. consider the exponential function § 2 d ¦ E 7 r ' ¦ ¦ d ³ 2 ' ± ¦ ¦ C 2 " § ¦ d y % r d 2 ÷ Ñ x ¯ § d W % w g ¯ d r 2 The above result tells us that ¥ ¤ 2 2 ' 2 g The above example exhibits a remarkable property of the exponential function. 6 u 2 u 6 f 7 ¥ 2 2 r ÷ Ñ W 2 ¨ ¨ 6 u 2 u 6 f for . Explain why the conclusion of Theorem 6. where .5 1.9 As a ﬁrst example of the use of Theorem 6. namely that the function and its derivative are identical. We still have not solved the equation ' " % " E % C 1 f E ( C 1 ! % " % % % % f E ( C ð 1 " E % C ð { i d ÷ ¯ ¦ ¦ ¦ d d ¯ ¯ § ' ' 2 d 2 d d ¯ ¯ d g g 2 " d E d 2 " 2 E C 2 y d d C x y " % w x w E 2 C ð y x w . In particular.94 SERIES AND THE EXPONENTIAL FUNCTION EXAMPLE 6. Find the radius of convergence of the series d ´ 2 a d 2 ` ° 2 U d § 2 ' d 2 g d 2 E 2 C What is the derivative of ? Justify your answer.
ó Î ò 7. 2 ò . Show that " u 2 u " f 2 d 2 d " 2 " f " Hence show that " u 2 u " f 7 ¦ 2 ' d 2 g d " " and E 2 . ó Î ò ¦ ¦ ¦ d ¯ ¯ E d 2 " . Explain why the functions ´ 2 ´ a ` ¦ 2 2 a a 7 ` ¦ 2 2 ¥ ¥ ¦ d Ó d Ó ¦ d f f f 2 ¦ ¯ ¯ ³ ° f § d " ¯ ¯ ³ ³ r ° U 2 2 d r 2 r d U 2 2 2 U Ó " Ó § d C d C d d d 2 d ¯ ¯ § ¯ ¯ § § ' g 2 2 d ' E ' g 2 2 and ¦ ¦ ¦ d ¯ 2 f " E 2 C 5. The functions and are deﬁned in terms of the exponential function by 2 C ð " u 2 E u 2 " f f is differentiable for . Show that the functions % { d ð ð { and of the previous exercise both satisfy the differential equation h g 2 are differentiable for all . Explain why the function 6. ó Î ò 2 . ò õ ô Show that ¦ ¦ ¦ d ´ d 2 2 . What are the derivatives of and ? ¦ ¦ ¦ d ¯ f 2 E 2 C " C . õ ò ô õ ô 2 .DIFFERENTIATION OF POWER SERIES 95 3. Show that satisﬁes the differential equation ¦ ¦ ¦ d f 2 E 2 C 4.
For example. the chain rule can be used to differentiate classes of functions other than those we have already considered. that is 7 " d 2 2 2 C E [ 2 . C ò õ ô 2 . One example is the identity O N M d r O N M E d r C O N M We shall prove this and other identities in the next section. We now deal with another way of combining two functions and to get what is called the composition of these two functions. ó Î ò E 2 . We shall meet them again in Chap 2 .4 The composition of two functions Let and be functions such that the domain of includes the range of . ó Î ò C ¨ 6. which we have not yet encountered. In Chapter 2.5. ó Î ò The functions ter 11. The composition is obtained by applying the two functions in succession. products and quotients of functions. [ DEFINITION 6. ò õ ô E 2 . In this section we shall develop a new rule for differentiation called the chain rule.96 SERIES AND THE EXPONENTIAL FUNCTION Hence show that 2 . but the proof requires a new property of the derivative. As well as enabling us to prove identities for the exponential function.6 THE CHAIN RULE The exponential function satisﬁes a number of algebraic identities which can be used to simplify computations. The two function machines and acting in sequence are replaced by a single machine . The composition of and is deﬁned by ç E 2 C æ E 2 C E [ C " d " 2 E E 2 C C [ E and we denote this function by the symbol . ò õ ô C 2 2 ¨ ¨ ¨ . we encountered sums. if 7 " 2 E 2 C 7 " d 2 E 2 C then " d " 2 E " d 2 C E E 2 C C In this way we have a new function deﬁned by the rule " [ for The process of composition is shown diagrammatically in Figure 6. and are hyperbolic functions. multiples.
. THEOREM 6. then the composition + ð +  is differentiable E  + Y ® ¡ % The domain of is . Solution. but if we take either of the points is in the domain of . together with their 5 5 2 3 _ 3 ` Ê 5 5 2 3 ` _ 3 ` Ê 5 _ 2 a 3 _ ` 2 Y + _ 2 has .10 Let and let domains. d É É Y + + Y + [ is given by É Y + [ Y + The domain of is . In order for has to be in the domain of . so that the domain of will be . Determine the functions and  +  Y [ + We also have to be careful about domains and ranges. The function + Y . In the above case. is deﬁned by The function [ Y [ Y Y + + Y [ B + + ~ ¡ + Y h [ h ¡ g h We now state the chain rule for the derivative of a composite function.5: The compostion of two functions [ Y Y [ In general and are different functions. then Y [ [ Y Ê Ê Ê B B B + + É É ð  B + ~  [ + Y EXAMPLE 6. so that the domain of B h Y [ h Y . For any other point the number will be in the domain of .4 The chain rule If is differentiable at and is differentiable at at and Y + ð [ Y + Y [ Y .THE CHAIN RULE 97 Figure 6. so this point must be excluded from the domain of . which is not . to be in the domain of and + Y Y to be welldeﬁned.
then " 2 2 and Consequently. If we want the derivative of . in effect. this result can be checked by using the quotient rule for 7 E E " " 2 d " g d 2 2 2 C f C f f E 2 C ð ~ " E 2 C . EXAMPLE 6. so that " d " 2 E E 2 C C { ~ " E 2 EXAMPLE 6.11 and If C " " d d 2 2 E E 2 2 C C . then i E " d 2 C ~ " As an exercise. the difﬁculty of the calculation is transferred to the proof of the theorem. The theorem is extremely useful. It is easier to understand and work with. Commonly. E " d 2 C ð E E 2 C C ð 2 g E 2 C ð and hence E 2 C ð E E 2 C C ð E 2 C ð E [ C There is an alternative notation for the chain rule.98 SERIES AND THE EXPONENTIAL FUNCTION This is a hard result to prove properly and we shall not attempt to do so.12 and If . but it lacks precision in certain cases. the hardest theorems to prove are the most useful in practice because. . then we put and and the rule is E 2 C ü E E 2 C C ü 2 ¨ ¨ { ü ¨ ¨ { 2 ¨ ¨ E ü C E E 2 C C { Then 2 g § " f { ¨ E " 2 d g " ü 2 C ü f { 2 ¨ " d 2 ü Put .
6 [ f E E 2 C 2 2 In exercises 1–6 you are given two functions . The following theorem summarises the main ones. and . ° E § 2 d " C E 2 C E 2 d " C E 2 C 7. ﬁnd the compositions and Differentiate each of the functions in Exercises 7–10.5 is to consider the function deﬁned on by where r and are any real numbers. (6. E C O N M E r C O The trick to deriving the properties of the exponential function in Therorem 6. 2 d " E 2 C 7 2 E 2 C and . ° ¨ d M d r d d C r E E O C " E N 2 2 M f ¨ ¨ C E O f f d N M r C d d O E N r r C C M 2 C C O O 2 N N ¨ M M ¨ E E E 2 2 2 C C C O O O E E N N N r E C % M M M C E " r O O C N N d " O M M N h ) M g % E d E w § ) f r E C 2 f r r ¨ C C C O O O E N N N ) M M M C 9. " 3 3 3 [ 2. 8. we have E E 2 f d r C O N M E 2 C O N M C ¨ This result holds for all real numbers . THEOREM 6. 4. It is an important result that a partial converse is also true. ( 6. ' d ü E ü C 7 ) E ) C 2 E 2 C 7 2 d 2 E 2 C 3. r E 2 C 7 § 2 E 2 C U f ) E ) C 7 ) " d g E ) C 5.7 PROPERTIES OF THE EXPONENTIAL FUNCTION As we have already mentioned. the exponential function satisﬁes a number of algebraic identities which can be used to simplify computations. In each case.PROPERTIES OF THE EXPONENTIAL FUNCTION 99 EXERCISES 6.13) .5 For any real numbers r N M E d r C O N M 3 and . " d 2 E 2 C 7 ' f 2 E 2 C 1. Using ﬁrst the product rule and then the chain rule. 2 r E 2 f E 2 d f r E C 2 O d N f M r C E Ê O 2 d C " N M O r C E N d E M O 7 2 N C E d M É 2 O E E N 2 2 f M C C E O f 2 N ¨ C 10. We know that if a function is constant then its derivative is zero. the exponential function satisﬁes the following identities. 6.
The proof of the ﬁnal identity of Theorem 6. The obvious value to try is . then as can be seen from the series deﬁning . . we shall make use of it in equation (6. The number is an irrational number and it plays a key role in the power series give applications of calculus to the real world. Calculations with the . so to ﬁnd . .14). We ﬁrst deﬁne a number by letting .13) and in many later situations. The properties of the exponential function in Theorem 6. % r which is the ﬁrst identity in Therorem 6. Thus for all . together with the result ¥ Ó " Ó E g C " y x w E g ° f Ó C y x w E U C y x w § Ó E ' C y x w and similarly Also . E % C y x w ¤ Ó E C 7 ° y ¥ x Ó w E U f C y x w § ¥ Ó E ' f C y x w and similarly . we have the following. We can also extract the second identity in Theorem 6. E E " C 2 C y y x Ó w x w E Ó " Ó C y x w E " C y x Ó w 2 E " Ó d % " C w y x E w 2 C y x E w g C y ! g x % w ! " w a % E g 2 w C y E x 2 Ó w f C y x w % 2 for all integers . since we know . If we put we ﬁnd 7 " E % C y x w E 2 f C y x w E 2 C y x w % 7 E d r C y x w E C y x w E r C y x w r 2 Finally.100 SERIES AND THE EXPONENTIAL FUNCTION This result is very important and surprisingly difﬁcult to prove and we shall not do it here. (possibly ). These results. Since and are deﬁned for some constant . .5.5 enable us to develop an alternative way of looking at this function.14) and r This is true for all values of .5 implies that is left to you as an exercise. we substitute any value of and solve for in terms of and . this equation implies that ð ¤ ¤ E 2 f ¤ d r E C 2 y C x w E 2 C y x w h so that E 2 C " y x w E 2 f C y x w If . Thus everywhere on .. Nevertheless.5 from equation (6. Once we have deﬁned . let in this last equation to get E d r C y x w E 2 f d r C y x w E 2 C y x w E d r C y x w ¤ so that and hence (6. then the function is .6 If the derivative of a function is zero on an open interval constant on . This last equation then . This gives ¤ 2 " E 7 % ¤ C y x w E d r C y x w E % C ¤ y x % w 2 2 h THEOREM 6. show that " .
EXAMPLE 6. when simply deﬁne by letting .14 and Find the derivatives of Solution. We have 7 ç E C O N M æ Ó E d C O N M " O N M Ó Ó so that To calculate and hence . . then we ﬁnd . We then have E 2 C O O N N M M Ó E C O N M k Ó i Ó k Ó 7 E C O N M h j E g Ó C h j g Ó i Ó EXAMPLE 6.13 Calculate to 4 decimal places.PROPERTIES OF THE EXPONENTIAL FUNCTION 101 How would we calculate the square root of the number ? We have as yet given no rule for calculating the square root of any number which is not an exact square. Solution. so that E 2 f C ¤ ¯ E 2 " f C ¤ ¤ The error h g 2 for all . we . We use the chain rule. we use the power series for the exponential function. together with the result M ¨ E 2 r C O E N M 2 r f E C E O 2 N r M C O N M C E 2 ¨ r C O N M E 2 E C 2 O r N g C M a O ! N U M ` E r 2 " C ð O 2 ¯ N r ` " 2 g ¨ d ¦ ¦ ¦ ' d % % " g % % d % " u n ´ Ó 2 ¨ g ~ " 2 If we put . We have ¦ ¦ ¦ d ¯ § ' 2 d ¯ g 2 d 2 d " E 2 C O N M Ó to 4 decimal places. Finally. but we can now give a procedure for calculating . We have 1 d ¦ ¦ ¦ d 1 1 O N M ~ E 1 1 C O N M Ó g Ó E C O N M Ó E C O N M terms h h Hence so that we can calculate by using the series for . where and are integers. We can similarly deal with any rational power . satisﬁes ÿ B ÿ A @ Ê h 1 E E É ? C þ O O N N M M þ C is an irrational number.
we found the differential equation for the air pressure problem to be (6. accurate to 2 % E 2 C ð ¥ % " Ó ~ " 10. Find an example of a function for which is not a constant. let us try to make use of some of it. Using the series deﬁnition of the exponential function.102 SERIES AND THE EXPONENTIAL FUNCTION Similarly E 2 r f C y x w r f E 2 r f C 2 ¨ ¨ E 2 r f C y x w E E 2 r f C y x w C 2 ¨ ¨ Another way of writing these results is (6. ³ Ó 11.0000000001. ° E § Ó f " C E 2 C 4.15) ß 7 ß Ó r ß Ó ¨ EXERCISES 6. E ¦ ¥ Ó d ¦ for all in the domain of . ﬁnd an approximation for 3 decimal places. ¥ Ó r f ß ¥ Ó 2 ¨ 2 ¨ ¨ Ó C E ) C (6. Ó 8. Equations of this form are applicable to problems other than the air pressure problem. where has no particular physical interpretation. o ¥ Ó { § ¥ Ó { 1. using the series expansion for the accurate to 9. Ó 2 d " E 2 C Ú Ú o Ó 2 g E E 2 2 C C 7. This means that your maximum error must be less than 0. to 3 decimal place accuracy. 3. 6. We write with . Can you reconcile this result with the statement of Theorem 6.7 Differentiate the functions given in Exercises 1–8. 2. The differential equation to be solved is (6.1.0005.6? 6. but for which 12. Ó § 2 { " " f d Ó Ó E 2 C 5. but since we have now developed an impressive amount of mathematics in deﬁning the exponential function. so let us emphasize this generality by writing . Find an approximate value of exponential function. so let us replace them by symbols for ease of working. How many terms of the exponential series would we need in order to ﬁnd ? 0.17) ' " % " E % C 1 7 ' " % % f E ( C 1 ! % " % % % % f E ( C ð 1 The constants are awkward to deal with.8 SOLUTION OF THE AIR PRESSURE PROBLEM In Section 6.16) accurate to .18) E 2 C { ¤ E % C { 7 ' " % { " r d ð ¤ { ' " % % f ! % " % % % % r 7 ¤ E % C 1 7 d E ( C 1 r f E ( C ð 1 We could solve this equation by a power series method. and . that is.
In summary we have the following result. If we replace the constant by a function right hand side comes out to rather than zero.18) by working through several steps. which shows that is a solution of the differential equation. { r f ß E ¥ 2 Ó C ð { { ß ¥ ¨ E % C { E ß 2 ¥ C ß { ¥ Ó " { E 2 C { ¨ This means that 7 £ E 2 ß C ¥ Ó { for some constant and so ß ¥ Ó £ E 2 C { This result shows that every solution of must have this form. then it follows that . { E % C { % { r d ð { % { r d ð { a particular We solve equation (6. First we solve the case when that is. so that ß ¥ { Ó { % £ { { r d ð { { E % C { The initial value problem (6.19) ¤ . E 2 C p % { £ r d ð { ß ¥ Ó { 1. The idea behind this is as follows: ß ¥ Ó { { is a solution of the initial value problem . so that .16). so we have to ask whether the solution is the only is some other solution. We have reached an important result which is worth some discussion. we solve the equation % { r d ð ç E ß ¥ Ó Ê C E E r f C Ó E C 2 2 C ¨ { E { 2 C f E E % { E C 2 { f C E { 2 7 C r % ð f { C E ß 2 ¥ C Ó Ó { { É æ £ r ß ß d " E ¥ ¥ 2 Ó Ó C % ð { Ê ß ¥ Ó É 2 ¨ £ . since every solution has this form. satisﬁes . We call solution of . We call % ß ¥ Ó £ { r d { ð { the general solution of the equation . % E 2 C (6. by the quotient rule. . Sometimes equations have more than one solution.20) has the solution Next we move on to the more general equation { r d ð { ß ¥ Ó { { E 2 C p . we may be able to choose 2. To solve the equation we make use of At this stage we will ignore the initial condition the result (6. so that the E 2 C p The strategy here is to replace the constant £ by a function . If we know the initial condition . solution of equation (6. Suppose that Then.SOLUTION OF THE AIR PRESSURE PROBLEM 103 .19).
The idea is to solve a simpler problem and then replace the constants by functions in a trial solution for the more complex problem.17) gives ' ' f ' E " ( % C " 1 ¤ 1 ' " % % f ! % " % % as the solution to the air pressure problem. In computing the exponential terms we could use the power series expansions. using % % .17) has the same form as equation (6. where is a constant.1.104 SERIES AND THE EXPONENTIAL FUNCTION This is an instance of a general procedure for solving differential equations we will encounter again later. in equation (6. If we substitute into the equation we get so that and hence ß ß ¥ Ó E 2 C ð p 7 ß ¥ Ó E 2 C p r d ß ¥ Ó E 2 C p r f ß ¥ Ó E 2 C ð p 7 { r d ð { ß r £ ß ¥ q µ Ó Ó Ê { ¤ ß ß ß ¥ ¥ ¤ d ¥ Ó r Ó ¥ Ó E ~ d Ó Ê Ê ß % ß ¥ Ó E ß { f d " r Ó f ¤ f " 2 É r C r 7 ¤ E d { É p 2 { É % C £ r ð E g d ß " d 2 { d p C { This is easily checked to be a solution of use of the initial condition . It is not difﬁcult to verify.21). If we now make and so This is another result worth summarising: The initial value problem (6. Thus the solution we are seeking has the form p E 2 C p We are now looking for a function the chain rule. It is known as the method of variation of parameters or variation of constants . We can check that this will reproduce the results of Table 6. but almost any calculator will have the calculation hardwired into it.21) has the solution r r { C r d ð { The air pressure equation (6. so that we can evaluate the exponentials at ¤ Ó r r ¥ Ó E 2 C { p { ¤ is such a function. Using the values . that whose derivative is Ó . we have ð { ¤ E % C { for any value of the constant . .
We have r E % C 2 ) ¶ § ) g § 2 f 2 ¶ 2 ¶ % % ) " ¶ ~ 2 ) ¶ d ) 7 ) 2 EXERCISES 6. The solution is ) Using the Leibniz approach to differentiation (page 46).20) with E ) C 2 . Set up and solve the differential equation for the amount of salt in the tank at time . Use the exponential growth equation in exercises 1–4. where is assumed to be small. EXAMPLE 6. we let becomes % 2 r f ) 2 ¨ ¨ ) and this approximation becomes increasingly accurate as we let equation in the form % 2 % g " f ) 2 ¶ ¶ approach zero. © A simple model of population growth assumes that the population tional to . the above differential equation (possibly with ) has many other applications. (But remember. including compound interest. The big payoff.) The unifying thread of this chapter has been the problem of ﬁnding an equation for the variation of air pressure with respect to height and we have developed an impressive amount of mathematics to reach the solution. Others appear in Exercises 6. % u ± ± % r ³ j ~ ¦ " ¥ Ó r r This is of the same form as equation (6. is not that we can now solve the air pressure problem. One of these is given in the example below. Then the initial condition is and .8 The exercises in this section can (and should) be done without the use of logarithms. Let be the amount of salt in the tank at time . This is often called the exponential or natural growth equation. That is 7 © ± ) © ¨ ¨ © where is a positive constant. which we consider in Chapter 9. radioactive decay and drug elimination from the bloodstream. it is that the mathematics we have developed is useful and applicable in a wide variety of seemingly unrelated problems.15 A tank initially contains 100 litres of water in which 5 kilograms of salt is dissolved. Rather. The mixture in the tank is pumped out at the same rate. Pure water is pumped into the tank at a rate of 2 litres per second. Solution. Let be change in the amount of salt in the tank during the the amount of salt per litre is time interval . It applies if birth and death rates are constant. As well as representing population growth. We rearrange this approach zero and the equation grows at a rate which is propor ) 2 ¶ ¶ % % " .8. however.SOLUTION OF THE AIR PRESSURE PROBLEM 105 the touch of a button. the designers of the calculator needed to have access to much of the theory of this chapter.
What will the temperature be after 2 hours? Assume that the room temperature is 25 C. A city had a population of 25 000 in 1980 and a population of 30 000 in 1990. by in the . Find the equation of the curve. y y y r ~ f { ü r ~ f { (b) Put solution. ) ¨ in the following way. Suppose that sodium pentobarbitol is eliminated from the dog’s bloodstream at a rate proportional to the amount of the drug present. How much of the drug should be administered in order to anesthetise a 25 kg dog for 2 hours? E { ð r { U 5. Suppose there are 1000 bacteria present initially and 1600 after 1 hour. and that half the drug is eliminated after 5 hours. the temperature of the cake is 90 C. t f t v Solve this equation and replace tation of ? v by in the solution. What will the population be in the year 2000? 2. t f t v (b) Put (a) Write the differential equation as 7. and show that the equation becomes © 4. The number of bacteria in a certain culture increases at a rate proportional to the number present. Show that t 7 E t f t C ± f t ¨ E ) C t t t Solve this equation in the following two ways: (a) Use equation (6.106 SERIES AND THE EXPONENTIAL FUNCTION 1. Solve the initial value problem . What is the physical interpre v ± f ) v ¨ ¨ % w ± where is a constant. The drug sodium pentobarbitol is used for anesthetising dogs. Let be the temperature of a body. solve the resulting initial value problem and replace Ê ¤ r f E % { C É { r f d { f ð { 6.21) in the text. Newton’s law of cooling states that the time rate of change of the temperature of a body is proportional to the difference between and the temperature of the surrounding medium. How many will be present after 3 hours? 3. After 30 min. ü . The dog is anesthetised as long as its bloodstream contains at least 45 mg of the drug per kilogram of the dog’s body weight. A curve passes through the point point is three times the value of the 7 % C © and has the property that the slope of the curve at any coordinate of . A cake is removed from an oven at a temperature of 150 C and left to cool.
1. . The mixture in the tank is pumped out at the same rate. 9.SOLUTION OF THE AIR PRESSURE PROBLEM 107 8. The average daily outﬂow of well mixed water from the lake is equal to the average inﬂow. Suggest reasons why this does not happen for the above solution. Water containing salt at a concentration of 0. check that the solution to the air pressure problem obtained above will closely reproduce the values in Table 6. A lake has a volume of 1 billion cubic metres of water.25%. The average daily inﬂow of water into the lake is 1 million cubic metres of water. Using Mathematica. with an average pollutant concentration of 0. Consequently. What will be the concentration of pollutant after 5 years? 10. The atmosphere only extends a ﬁnite distance above the surface of the earth. Show that the differential equation for the amount of salt in the tank at time is ) 2 % 2 q µ ( R g ¥ Ó ' f ' r " " % d % ) 2 g ¨ ¨ " f { Hence ﬁnd the amount of salt in the tank after 10 min.05%. A tank initially contains 100 litres of pure water.1 kg/litre is pumped into the tank at a rate of 5 litre/min. with an initial pollutant concentration of 0. we would expect the pressure to approach zero as the height increases indeﬁnitely.
1).1) { t g f ) ¨ . Since the body does not move horizontally. then its mass per unit length is . a problem we will consider in detail in Chapter 8. If the total and the string has length . The string vibrates because it is under tension and this tension exerts a force and respectively and we let on the body. but this requires mathematics beyond that which we will develop in this book. we shall make a simple approximation and assume that the mass of the string is concentrated in a single body at the center of the string (Figure 7. the two horizontal components and must be equal and opposite to each other and we can ignore them as far as the vertical motion is concerned. Instead. 7. a horizontal component and a vertical component. The tension in each half of the string can be split into two parts. The frequency we are considering here is the frequency of vibration when there are no outside inﬂuences.1 VIBRATING STRINGS AND CABLES The problem to be considered is the vibration of a cable or string with ﬁxed end points. We denote the tension in each half of the string by be the displacement of the centre of the string from the equilibrium position. It follows that the tensions and are equal and hence that the two vertical components are equal. To make it vibrate at a different frequency is more difﬁcult. we have to consider each part of it. Let be the common value of and and let be the value of the vertical component. then the system will readily do so. but we can still make progress by making some assumptions and simpliﬁcations.CHAPTER 7 TRIGONOMETRIC FUNCTIONS In this chapter we consider the vibrating cables of the Anzac Bridge. The numerical information we want to derive is the frequency at which the cable vibrates. Newton’s second law of motion implies that { t g { t t t t t 2 ¨ t q t q t 2 S ~ t t S (7. This is called the natural frequency. On a large scale this could apply to the Anzac Bridge. A full analysis of the problem would take us well beyond the mathematics in this book. The mass is spread out along the string. while on a small scale it could apply to a musical instrument. Let us consider a vibrating string or cable. The two vertical components add up to give a total force of on the body. If some outside inﬂuence tries to make a system vibrate with its natural frequency. To analyse the mass is motion of the string. This is a task for a later subject. such as a violin or guitar.
1.1) gives us 2 t U 2 ¨ or 2 t U f 2 ¨ 2 We shall assume that the error in this approximation is negligible for small 2 t g { t 7 2 7 S S g t 7 S S ~ U S S g f S f { E ) t t g ) ¨ ~ ¨ S C d 2 · 2 If is small. 2 t g Using this result in equation (7. we have 2 { t so for small .1: A vibrating string or cable Using similar triangles in Figure 7.110 TRIGONOMETRIC FUNCTIONS Figure 7. 2 { t 2 that is. then and write 2 g ~ S { t E g ~ S C t q d t ~ 2 } · { t q g t t t { t g ~ S .
Polynomial functions will fail for the same reasons that they failed in Chapter 6. as a trial solution. we arrived at the initial value problem (7.3) C 2 instead of from You may be puzzled by the fact that the sum now goes from is easy to see if we ﬁrst expand the ﬁrst few terms of the sum (7. a function deﬁned by a power series: (7.2) to get Ñ W r The problem is to determine the coefﬁcients 2 ¨ " f 2 2 ± § C ) ± ¨ d d ÷ ) ) ¨ ¨ . using Theorem 6.2.4) . The reason .2 TRIGONOMETRIC FUNCTIONS In the previous section.3) into equation (7. which leads to the initial value problem 2 2 ¨ E % C ð 2 2 E % C 2 2 and with velocity The differential equation in this problem is known as a second order differential equation .TRIGONOMETRIC FUNCTIONS 111 We shall write this as 7 % 2 d 2 ) ¨ ¨ where t % ± 2 ¥ ) r E 7 7 " % ¦ ¦ f ¦ 2 2 ± d C ) ± ° g ) E E r ° % % ÷ Ñ C C ¥ r W ð ð ÷ Ñ ± ) W 2 2 d 7 7 § r ¦ ) E ) ¦ § 2 2 " ¦ r d r f d d % g E E ¥ ± ÷ Ñ % % C ) W ) C C ) ° ± 2 2 r ) r r ' 7 7 ÷ E E Ñ % % d W ) We shall assume that the motion commences at with the mass displaced by . since the derivative appearing in the equation is of second order.2) 2 ¨ To solve this equation.3) to get ) r d r E ) C 2 and then take the second derivative U d ) r g ' d r g E ) C ð ð 2 Since the derivative of a constant is zero. We will discuss solutions of this equation in the next section. we have (7. one term drops out for each differentiation. Let us use. Thus and .4) and (7. we use the method of inﬁnite series as developed in Chapter 6. We substitute equations (7. On differentiating twice. 7.
( For the evennumbered coefﬁcients. æ Ñ W which shows that (7.5) relates each coefﬁcient to the second one before it. Since Equation (7.6) and (7. . the evennumbered coefﬁcients ) and the oddnumbered coefﬁcients ( ) will be determined separately. replacing by in the ﬁrst sum. we have r ³ r ° r § r r r and in general 7 ' 7 g 7 " ± 7 r ¯ E Substitution of equations (7.3) gives ° E ) C E ) C 7 ¯ ¯ a ` f Ê ¦ r ` ¦ ¦ a ` 7 d f 7 ' ³ 7 ' E ¯ g 7 ) 7 r g ´ " 7 C r " d ± § ± E ¯ 7 ) ' ¯ ¯ 7 C r U Ê % ¦ f ¦ ¦ ) f r É ' U ¯ r U r U d d f " E Ú " E E d d " ± r f ³ ° ± g E C ¯ f r C g E C g C " d f ± 7 C " Ú E r ¯ ¯ É g g ' r ) If we equate coefﬁcients of d ± C on both sides of the equation for 7 % % % ç ) ) ç ) r r r ÷ Ñ W d d ) Ú d Ú r ) E r " Ú E " d r E ± d C " E ± C g d E g ± d C E ± d C g ± C d ÷ ± C ÷ ÷ Ñ 9 g r f d E ± ) C " g 2 ± g ' that is. we ﬁnd f (7.8) .6) § r (7. for the oddnumbered coefﬁcients.5) is called a recurrence relation . we have 7 r f ° r f r 7 ³ r 7 7 § ° r 7 r r r f r 7 r f r 7 f r 7 ° r r 7 r 7 r and in general 7 ' 7 g 7 " ± 7 r ¯ Similarly.112 TRIGONOMETRIC FUNCTIONS Hence æ Ñ W that is.5) r Ú r E " d ± C r E g d ± C f Ú r in terms of .7) into equation (7.7) (7. it gives the coefﬁcient equation (7.
2) (7.8) with respect to and then put .10) .9) Of course.1 The sine and cosine functions . we differentiate equation (7. We deﬁne the sine and cosine functions as follows: DEFINITION 7. Thus has domain . we have to verify that the series in Deﬁnition 7. Finding is easy. Using Deﬁnition 7. equation (7.2 1. the series converges by the ratio test and the radius of convergence is inﬁnite.11) r ) ò õ ô ) ó Î ò r d h ) ) ò ò õ h õ ô ô h r g ) E ) C 2 ) ó Î ò h g ) The arbitrary constants and can be found using the initial conditions. we deﬁne For all Ú E " f C Ñ 9 ´ ) ³ ) § ) h g ) (7. To ﬁnd . Use the ratio test to show that the series (7. We put in equation (7.2.10) converges for all real numbers . so that .8) and ﬁnd . although the connection will not be obvious at this stage. A similar calculation shows that the series for converges for all and so the domain of is . We can do this by means of the ratio test. The ﬁnal solution to the initial value problem (7.12) E ¦ ¦ ¦ d ú ú ú ú ú ¯ U ) ¯ d ) ¯ E E " ¯ " ) ' ¯ f E d E ± " ± ± g f g f C g C C C 7 C E ÷ ÷ " Ñ W f ) d ± ¯ ¦ ¦ ) E  E g ¯ ¦ ¦ ¦ C " E ¦ ¦ ¦ E ) "  ¦ " d d d d ± ¯ ¯ d d ± g a ` ) ± C g ¯ g C f f C r ú ú ú ú ú ¯ ¯ ° r Ñ U Ñ ) Ï Ï f r Î Î Ð Ð Í Í d d % ¯ ¯ ¯ g ' ' ) ú ú ú ú f f d ) " ¥ ) r ú ú ú ú ) ) ò f ó Ñ Î C Ï õ ò Î Ð ô Í r E ) C ð 2 (7. They are functions which you may have already met.1 actually converge. We ﬁnd ° ) ³ ) § r r ³ ) r ° E § % C ) 2 r r % ) % ) ) EXERCISES 7.1.8) can now be written (7. This procedure is justiﬁed by Theorem 6. ) ) ó Î ò 2 d ) ò õ ô 2 2 r E ) C 2 r E % C ð 2 % ) Putting is thus gives .TRIGONOMETRIC FUNCTIONS 113 The two series in parentheses are extremely important. The general term of the sine series is Ú E " f C so that ¥ ¥ E " f C Ú E " f C r Since this limit is less than 1 for every .
we note that each term is obtained from the previous one by multiplying it by an expression of the form ) f ) and then calculate the terms 7 ¯ E U ° Ú d ± ) g C 7 ¯ E g Ú d ± ) g C until we ﬁnd one whose magnitude is less than the required error. To see that the terms will always start to decrease eventually. The centre of the string is displaced a distance of 5 cm from its equilibrium position and then released. the mass per unit length of the string is 0. The ends of a string under tension are ﬁxed at two points and .3 MORE ON THE SINE AND COSINE FUNCTIONS Deﬁnition 7. we use .114 TRIGONOMETRIC FUNCTIONS 2. Such a series is called an alternating series. but as long as we take enough terms. if ) ò õ ô ) ó Î ò ´ U ³ U § U ) U 2 ò õ ô ) 2 ó Î ò ) If the value of is large. as long as the terms are decreasing in magnitude. This differential equation arises in all oscillation problems. These functions have been introduced because they provide solutions for the differential equation (7. Thus to calculate the following rule: choose so that " u E g d ± g C E ) " d ± g C ± ± ¤ £ ¤ ¦ ¦ ¦ d ¦ ¦ £ ° ¦ § d E ' ¯ g a f d ³ f µ ± ¯ ¯ E g ! r C E d g " C d § ¯ d % ' " ± g C f f U U U ó Î ò . is less than the ﬁrst term omitted.2). The distance is 2 m.1 kg and the tension in the string is 2 N.1 provides the basis for all calculations with and . For example. We shall later relate these trigonometric functions to the properties of triangles. Suppose this term is ¤ ) Then 7 ¯ ¥ E ¤ g ) f ¤ E " g C f C d ¦ ¦ ¦ f ¯ ° U ) d ¯ g ) f " ) ò õ ô ) ò õ ô For large enough the absolute value of this term will be less than 1. Find the value of the constant in equation (7. 7. there may be many increasing terms before a decrease begins. The power series deﬁnitions enable function values to be calculated to any desired accuracy and so the functions are welldeﬁned. In the case at hand the terms of either series may initially increase for some values of . The error. These properties are used to deﬁne trigonometric functions in elementary mathematics. but it should be noted that the deﬁnitions in this form do not enable function values to be calculated to arbitrary accuracy. There is a simple rule for calculating the error term for such a series.12) and use this equation to ﬁnd the displacement of the centre of the cable after 2 seconds. we get to decrease. they will eventually start in the sine series. and each have the property that successive terms have opposite signs The series for for any value of . in absolute value.
1 Suppose we wish to evaluate with an error of less than . We ﬁnd % µ 7 C 7 so that % < H < B ¯ < 7 d ¦ ¦ ¦ ¯ ° 7 7 d ¯ C 7 B 7 ò õ ô E with an error of less than .0001. Since . We ﬁnd = ° C + Thus we omit all terms after the second to get % C + C ò õ ô with an error of less than 0.00007.MORE ON THE SINE AND COSINE FUNCTIONS 115 with an error of less than ¯ ¤ C 7 ò B 7 ¯ % % õ B C + all terms after the ﬁrst are decreasing and we calculate successive terms.  K C % H 7 d ¯ ¯ < C § § K C ¯ 7 + C ¯ d < C + ¯ 7 C + ¯ C < + ò õ ô EXAMPLE 7.2 To evaluate with an error of less than 0. we ﬁrst look at terms of the form 7 ò õ ô This gives M < This means we have to evaluate the terms % µ 7 until one of them is less than the allowable error. In terms of inequalities we have = K C ô K = It is common to write results such as this in the form = ® C ò õ ô EXAMPLE 7.
THEOREM 7.1 Derivatives of the sine and cosine functions For all . Similarly. and similarly. we ﬁnd t % ) 2 f 2 ) 2 ò õ ô ) ) " E ° ) 2 f C % ò ò õ ) õ ô ) ô ó Î ò E t d % ) C ¨ ó 2 Î ò ó % Î ò ó Î f ò ) ò õ E ô 2 f C E ó ) Î t ò ó Î d ò ) C ò õ ô while if we differentiate equation (7. but as we shall soon see there are properties of the functions which shorten them considerably. we reach the following important theorem. There is. we ﬁnd . E ) f d r C ó Î ò ) ó Î ò d E ) f d r C ò õ ô ) ò õ ô E ) C ð O N M r O N M E d r C E O ) N M This theorem can be proved by a similar method to that used to show We begin with the function E ) f d r C ó Î ò ) ò õ ô d E ) f d r C ò õ ô ) ó Î ò E ) C f d Ê 7 ¦ r ¦ ¦ C ¦ ¦ ò ó ¦ Î õ d ò ô d ¯ r ) ¯ a ò ó ` Î õ ò ) ô f f ó ¯ f Î f ¯ r ò E U ) ò f d õ d f ¯ ô ¯ ' r E g ò ) f d õ ò f ) ô õ r C É " ô C ó ) Î ) f E ò ¨ ¨ ¨ E ) ) ó d Î E ó Î ò ) r ò C ò C f ò õ ) õ ô ¨ C % ô ) ¨ . If we use the rule for differentiation of a power series to differentiate equation (7. We shall need some preliminary that results before we can demonstrate this. however.2).1.116 TRIGONOMETRIC FUNCTIONS These calculations can get very lengthy as the argument of the function gets larger. using the product rule and the chain rule. we ﬁnd ´ ) ³ ) ) § ) ¨ By comparing these equations with Deﬁnition 7. These properties have to be deduced from the power series deﬁnitions or from the original differential equation (7. THEOREM 7.10) with respect to . If we put series for . . This means that there is a particular number such and for all values of .1. . . The ﬁrst result we need is contained in Deﬁnition 7.2 Identities for the sine and cosine functions For all real numbers and we have ó Î ò r ò õ ô d ò õ ô r ó Î ò E d r C ó Î ò r Then. we have ) ò õ ô E ) ó Î ò C ¨ h g ) The next theorem states two important properties of the sine and cosine functions. in the Some properties are easily deduced from the power series deﬁnitions. .9) with respect to . a fundamental we see that property of the sine and cosine functions which is not at all obvious from either the power series or the differential equation and that is their periodicity . If we replace by in the sine series.
In Exercises 5–12. g E ò õ ô f ó Î ò C d E ò õ ô d ó Î ò C 1. 2. we deduce the following result. enables us to deduce that ) ó Î ò ) r ó Î ò ó Î ò d ) f ò r õ ô ò ) õ ò ô õ ô r E g ) ò õ f ô ) C ò õ ô ) d ~ 2 ó ó 2 Î Î ò r If we put in the second identity in Theorem 7. Hence . ﬁnd the derivative of the given function. from Theorem 7.2 we can get the following double angle formulas by putting r r ò õ ô r ó Î ò g r g ó Î ò " ) ó Î ò d ) ò õ ô and r ó Î ò g f " r g ò õ ô All of these results have been obtained purely from the power series deﬁnitions of the functions. The proof of the second one is similar. where we have used the results . . Putting then gives the ﬁrst identity in Theorem 7.3 Identities for the sine and cosine functions For all real numbers and we have ó Î ò r ò õ ô f ò õ ô r ó Î ò E f r C ó Î ò r " ) % ó Î ò ò õ ô f % E ) % f C ó Î ó ò Î ò ) ò õ ô E ) f C ò õ ô E d r C ó Î ò ¤ % f ) r ) Hence Finally. we get E ) f ó d Î ò r C r ó ó Î Î ¤ ò ò ) ò d õ ô ò õ d ô E r ) ò f õ ô d E r C ¤ ò f õ ô r ) E C ) ò ó C Î õ ò ô ¤ % E ò ò ~ ) " C õ § ð ô f f Since . " " . . If we replace by in this theorem and use the results .3. 2 ò õ ô 2 ó Î ò f " 2 2 § ò ò õ õ ô ô d d 2 2 ó Î ó ò Î ò 4.MORE ON THE SINE AND COSINE FUNCTIONS 117 If we put we get . " f r ò õ ô g r g ò õ ô ~ ~ " g ò õ ô r ~ ó 2 Î ~ ò ó 2 ó Î Î ò ò d ò " õ r ô d d " ò " ~ õ ô using the result in this last equation.3 Prove the trigonometric identities in Exercises 1–4. 3. EXERCISES 7.2. we must have for some constant . THEOREM 7.
suspended from a ﬁxed point by a light inelastic string.118 TRIGONOMETRIC FUNCTIONS (c) Give an example of a pair of functions with these properties. 10. 7 ¥ Ó ò õ ô d ¥ Ó ó Î ò ¥ ¥ Ó Ó ó Î ò { f Ó ð ð f { ¥ Ó ¤ d Ó { 16. Find the value of ¥ E 2 C d E 2 C ° (a) Prove that . and are constants. Let and be differentiable functions such that and % 2 7 " E 2 C 2 ó Î ò ' f 2 ò õ ô E 2 C 11. a constant . ~ ¤ where and are arbitrary constants. A simple pendulum consists of mass as in the diagram. 6. 7 E d ) ) C ò õ ô £ E ) C ~ E ) C The angular displacement of the mass at time is given by ~ 17. Show that satisﬁes the differential equation % E r C " E r C r (b) Suppose there is a number such that and . 8. . ò õ ô Ó E C 2 ò õ ô 2 { 7. 2 E 2 C E 2 C ð E 2 C f E 2 C ð ¥ £ £ 13. ) ' ò õ ô ) g ó Î ò E ) C 2 ó Î ò E 2 C 5. % k 2 7 2 ó Î ò 12. for all . Show that satisﬁes the differential equation 2 ò õ ô Ó { f ð { g f ð ð { 2 ò õ ô Ó § f £ { 15. 2 g ò õ ô 2 f { 14. § E 2 ó Î ò d " v C E 2 C 2 ò 2 õ ô ó Î ò f " E 2 C 9. Show that satisﬁes the differential equation 2 g ó Î ò { U d ð ð { where .
Notice. At each instant the direction of the velocity is tangential to the circle (Figure 7. while is positive. we have In triangles  »   ©  Hence 7   2     0  0  and  so that  2 0   0  and {  You should convince yourself that and always have opposite signs. Let’s look at the magnitudes ﬁrst. they do not give any systematic method for calculating function values. but we have to be careful about the sign of these quantities. so that 2 0 2 0 0 { 0 { 0 f 0 0 f  0  % u 0 In the diagram shown. The number arises naturally in connection with trigonometric functions and here we can also make connections with the power series deﬁnitions.1 Trigonometric functions deﬁned by triangles In this section we connect the series deﬁnitions for the trigonometric functions with the more familiar deﬁnitions in terms of the ratios of various sides of a rightangled triangle. no matter where the particle may be on the circle. Suppose we have a particle moving around a circle of radius such that the velocity has constant magnitude . so this equation always applies.4.2). so . These functions can also be deﬁned by using the unit circle. Hence  £ »   © £ and  © £ 0 £ ¤ » 7. TRIANGLES. % d ð ð (a) Show that satisﬁes the differential equation ~ .(b) Show that can be written in the form 7 ) ó Î ò d ) ò õ ô r E ) C ~ You will already have met the sine and cosine functions in connection with triangles. We can get the magnitudes of and by using the fact that triangles and are similar. 7. Similarly. and .4 TRIANGLES. always has the same sign as . that while the triangle deﬁnitions may look simpler than the power series deﬁnitions. CIRCLES AND THE NUMBER 119 ~ ~ ~ % » w 0 { . but this is just a generalisation of the triangle deﬁnitions. CIRCLES AND THE NUMBER 2 ¤  0 0   { {   £ £ 0 0  » ¤      0    0  0  © £ £   © £ £ ¤ r where and are constants. The functions we have deﬁned in terms of power series are the same functions. parallel to the axis and a component parallel The velocity can be resolved into a component to the axis. but we have not yet shown this.
120 TRIGONOMETRIC FUNCTIONS Figure 7.14) to get 0 ð ð Finally we note that ) ) 0 arc ~ Hence { ~ ó Î ò ) % ó Î ò E % C ð f 2 If we assume that the particle starts from the point . Comparing the 0 ð { P 0 ð ¥ 0 £ 2 ¤ 2 { 0 f 0 0 2 ~ ð ~ { © 2 » ¨ and . then this last equation becomes (7.15) is and so two expressions we have found for gives the result ð 2 E % C 2 ) ò { õ ô ) ó Î 2 ò ð 2 % 2 ¥ £ ¥ d ð ð 2 ~ 0 If we write .2: A particle moving in a circle On the other hand.15) . With these initial . so (7. from the deﬁnition of velocity we know that 2 7 { 0 f ð 2 We differentiate equation (7.13) and make use of equation (7. then and conditions the solution of equation (7.13) (7.14) .
For historical reasons. We can show that is decreasing on the interval by showing that its derivative is negative on this interval. We know that and it can be veriﬁed as an exercise that . Thus satisfying g ~ " % % % ò õ g ô ~ ) ò ò õ õ ô ô % ) ò õ ô % ) ` " ) ò U õ % ô f g ò õ ô How do we calculate the number ? We begin by showing that is a decreasing function between 0 and 2.3: The graph of on the interval Á 1 " f 2 TRIANGLES. that is. the terms of the series for are all decreasing in magnitude and so ) g ó ~ Î ò g 7 % g 7 % § ` ) g ) f ) ) ) ó Î ò ) ò õ ô ) ó Î ò Thus is positive on which implies that is negative on . For ﬁxed . 7. Let us consider the smallest positive number least one number between 0 and 2 where . so will be uniquely deﬁned. this number is denoted by . Hence is decreasing on and there is only one zero.Similarly. CIRCLES AND THE NUMBER 121 " % . ) ò õ ô g 7 % ) ò õ ô ¨ This implies that its graph cannot cut the axis more than once in this interval. 2 ~ ò õ ô These last two equations show that the functions we have obtained as power series from the differential equation are indeed the familiar functions from elementary trigonometry. The cosine function is differentiable and so continuous and hence there will be at . We can crudely ﬁnd by trial and error by checking ) ò õ ô g 7 % g ~ E ) ò õ ô C E ) ¨ ~ ¨ C g 7 % g 7 % ) ó Î ò E ° g E f ° ) g f § ) g g % " " ³ g ) " f f % % f g g " " § ` C C ) % % ) ) f g g " " % ) w w w ) ó Î ò % g " ~ ³ ) with a maximum error of .2 The number Consider points where .3. we wish to show that the graph of on looks like the graph in Figure 7. Thus g 7 % ) ò õ ô { Figure 7.4.
we have " ) ó Î ò d ) ò õ ô g ~ " g ) ~ g ó Î 7 ò % % Many other results follow from the addition formulas. The calculations above show that 7 ! r " u g ~ u a r " so to two decimal places. % ó Î ò " f ò õ ô This last pair of identities demonstrates the periodic property of the sine and cosine functions. This is a very inefﬁcient method for ﬁnding . By considering the signs of the derivatives and the points where the derivatives are zero. To sketch the graphs of either of these functions. t E g d ) C ó Î ò E g ' d ) ` C ò ¨ õ ô g 7 g % ) ó Î ò { Figure 7.4: The graph of Á g ) ) ò ó õ ) ) Î g ô ò ó ó Î Î % f f " ò ò % % E E E E f ) ) ) ) g ò f f f d õ ! g ô r C ~ g g ò C C " õ C ó ó ò ô Î Î ò ò ò õ õ ô ô ! ) ) ) ) % ò ò ò ó % õ õ õ Î % ò ô ô % ô % " " E E E E f ) ) ) ) g a f f f d r ó Î g " ò . If changes sign between two values of . 7.122 TRIGONOMETRIC FUNCTIONS positive and negative values of . If we use in the identity we ﬁnd and since is positive on . then it must be zero for some number between these two values. we also know or . we arrive at the following graphs. there are a number of other cases where we can compute exact values of the sine or cosine functions. Once we know or .5 EXACT VALUES OF THE SINE AND COSINE FUNCTIONS As well as the result . we to on page 116 is in fact equal to need only sketch it in the interval and then repeat this section in both the positive and negative directions. We see that the number referred . We have ) ' % % f ` " ò õ ô ) a ò % õ ô % ) r ò " õ ò ô õ C ~ g g ò C C ó ô õ Î ò ò C ô g ò õ õ ó ) ~ ô ô Î ò ó Î ò ò f õ ô ) ó Î ò ) ò " U õ " ô ' g g ~ f ó Î ò and we can keep going until we reach the desired accuracy. but in principle we can calculate it as accurately as we wish.
4 Show that " g ` ó Î ò We have ' ó Î ò ÿ ` f g þ ó Î ò We expand both sides to get 7 ò õ ô ó Î ò g ó Î ò ò õ ô f ò õ ô ó Î ò so that ò õ ô ó Î ò g ò õ ô i " g ` % ó Î ò k ò õ ô As an exercise. .EXACT VALUES OF THE SINE AND COSINE FUNCTIONS ¨ 123 EXAMPLE 7. It follows that " ) ò õ ô d ) ó Î ò ` ` ` ` U " ó Î ò " g ` ` f Notice that both functions only have values between and because ) ò õ ô { Figure 7.3 Show that g " U ó Î ò We have ÿ U d U þ ò õ ô g ò õ ô % Hence g " U ó Î ò EXAMPLE 7. you should show that .5: The graph of Á g f " g " " ` f g f g f .
rather than their relation to triangles.124 TRIGONOMETRIC FUNCTIONS Table 7. 20. together with the addition Use the exact values of the sine and cosine functions given in Table 7. and formulas for sine and cosine to derive the following results. ô ó Î ò " f ò õ ô . ) ò õ ô ÿ ) f g þ ó Î ò ) ó Î ò ÿ ) f g þ ò õ 5. " ó . 19. " ' g U ô M ò " U r ó . ) ó Î ò E ) f C ó Î ò ) ò õ ô f E ) f C ò õ ô 7. 8. 12. 6. " g f ` a ó Î ò g U a ò õ ô 17. . 18. 14. 4. .1 to derive the results in Exercises 13–20. All of the results of this section have been derived purely from the power series deﬁnitions of the functions and these power series were obtained from the differential equation. ) ó Î ò E ) d U g C ó Î ò ) ò õ ô E ) d g C ò õ ô 11. EXERCISES 7. . Further exact values can be found using the periodicity of these functions or properties such as and . 16. % % ó . % % 1. The importance of these functions in calculus is their relation to oscillation problems. ' " ` ó ' ó 15. 10.5 " g ~ ó Î ò 2 ò õ ô E 2 f C ò õ ô 2 ó Î ò f E 2 f C ó Î ò Use the results .1 summarises important exact values for the sine and cosine functions. 3. " f ' g f g " f " g f % g ~ " g ò õ ô g " % % ' g ó 1 ) 0 " g g " ' g ' g g " " 0 g ) 1 ` r U ' ' g g ' U ` Î ò " % % ò ) ó õ Î ò ò ô õ ô " g ò õ 0 . 13. You may also need to use other properties of these functions.1: Exact values of the sine and cosine functions Table 7. We have not made use of the deﬁnitions of the trigonometric functions as ratios of sides of rightangled triangles. ô g ó Î ò 2. ) ó Î ò f E ) f g C ó Î ò ) ò õ ô E ) f g C ò õ ô 9. such as the addition formulas.
4 The function ) ô M ò ) ó . ó ) ) " ó . ) Î ò ô ó M Ê Î ò ) { ò ) ) ) ò cosec 7 Of these. the tangent function is the most important. These are the tangent. which are derived from the sine and cosine functions. " ) ó . E d r C ó g .OTHER TRIGONOMETRIC FUNCTIONS 125 7. Among these are the following. ) ò " õ ô ¨ " " g ' g ' ) f f f ô M ò ó ó . f 3 . r ) ) ò d ó ó Î õ r ò ô . Notice that is undeﬁned when . ó f f . It is obtained as follows. secant. THEOREM 7. The derivative of is also important.6: The graph of ) ô ò ô ) The expression cosec is often written as .6. deﬁned by ) % ) ò õ ô Á ) ò ó Î õ ò ô g ) . The trigonometric functions introduced in this section satisfy a large number of identities and have various properties which can be derived from the properties of the sine and cosine functions. when 7 g ~ r ® 7 g ~ ' ® 7 g ~ ® ) Figure 7. The graph of the tangent function is given in Figure 7. õ ô ) ó ) 7 . 7 . that is. C ) ¨ ¨ ) ó . cosecant and cotangent functions. d " 3 satisﬁes the identities ó ò d Î õ õ ò ô ) ) ô É ò ò " ) õ õ ¨ ¨ ô ô E ) ó . ) ó . mainly by American authors.6 OTHER TRIGONOMETRIC FUNCTIONS There are four other trigonometric functions.
Show that ó ó . 9. " E d r C ó . Assuming that the statement is correct. Show that the sequence ¡ ¤ 2 converge to? Test your answer by taking some speciﬁc examples for Õ ¤ 2 ¤ 2 ó Î ò Ô ¡ % ¤ 2 is continuous at . ﬁnd the derivative of the given function. Derive the identity 2. 2 . õ ô 2 f E 2 C 2 ¨ ¨ 2 ó 2 ô M ò E 2 ô M ò C 2 ¨ 3. Õ ¤ 2 ¤ 2 ó Î ò Ô ¡ ¤ 2 11.6 1. r d ó r . õ ô C 2 ¨ 5. cosec 2 f E 2 . cosec . d . Suppose you are told that the function 2 ó Î ò ¬ converges to 1. .126 TRIGONOMETRIC FUNCTIONS EXERCISES 7. " . ó f . In Exercises 6–9. ¨ 4. õ ô 2 ô M ò 2 ó . cosec cosec 2 2 ô M ò In the following three exercises. cosec and . { 2 ó . derive the given expressions for the derivatives of . what does the sequence 2 % % 2 k 2 2 { 7 2 7 " © E 2 C 2 U ° 8. Hence show that the function in the previous example is continuous. Let about the continuity of at 2 % 2 be a sequence converging to 0. { 10. { 6. 2 . ¨ cosec 7. Let be any sequence converging to 0. 2 g ó Î ò 2 ó .
will analyse the equation using the example of a body of mass Figure 8. where is the total force on the body. We hanging on a spring (Figure 8. There is the force due to gravity which pulls the mass down. Next there is the force in the spring itself.CHAPTER 8 OSCILLATION PROBLEMS 8. There are a number of contributions to the total force. automotive suspensions or electric circuits. whether these involve vibrating strings.1 SECOND ORDER LINEAR DIFFERENTIAL EQUATIONS The differential equation obtained for the oscillations of a stretched string occurs in almost all oscillation problems. while if the mass is pushed up the spring pushes it back down. so that û 7 E f 2 C ± f û { Equilibrium position 2 ) ¨ . Hooke’s law states that the spring force is proportional to the extension of the spring.1).1: A weight on a spring 2 Let be the unstretched length of the spring and let Newton’s second law of motion tells us that 7 û 2 ¨ be the distance from the suspension point. It is given by 7 û û û where is the acceleration of gravity. If the mass is pulled down the spring pulls it up.
128
OSCILLATION PROBLEMS
where is a positive constant. The body may experience frictional or damping forces due to air, immersion in a liquid or having a dashpot or shock absorber attached. Such forces are usually proportional to the velocity, but act in the opposite direction. The frictional force can written in the form
7 ) 2 ¨ ¨ f û ±
where is a positive constant. Finally there may be an external timedependent driving force which . The total force acting on the mass is can be of any form. We denote it by
E ) C d û d û d û E ) C û
and so the differential equation is
(8.1)
We will do a little work on this equation before attempting to solve it. First rewrite it as
E ) C 2 ¨
Next, put
± f f 2 {
We do this in order to simplify the later mathematics, but there is also a physical interpretation of which comes about as follows. If there is no external force and the particle is hanging at rest, then , and equation (8.1) gives us
{ % E ) C ð ð 2 % E ) C ð 2
This is the equilibrium position, so is the displacement from equilibrium. In order to substitute the expression for into equation (8.2) we need the derivatives of in terms of the derivatives of . These are
2 7 2 ¨ { f f 2 É ¨ { { ¨ {
This is the basic differential equation for forced damped oscillations. It is a second order equation and we shall ﬁnd solutions in terms of exponential and trigonometric functions. The equation is said to be linear because it contains only a sum of ﬁrst powers of derivatives. More precisely, a second order linear differential equation is one of the form
E ) C û { E ) C ¤ d ð { E ) C d ð ð { E ) C r
E
)
C
û
~
E
)
C
E
)
C
û
9
g
{
d
~
ð
{
9
g
d
ð
ð
{
~
±
Using these results and putting equation (8.2) as
,
and
E
)
C
)
¨
d
)
¨
)
2 ¨
¨ Ê f ± d E ± Ê f ± d 2 C ± f f 2 2 ) ) f ¨ ¨ ¨ 2
É ) { ± ) ¨ 2
)
¨
¨
¨
¨ d 2
) ¨ ¨
(8.2)
enables us to write (8.3)
SECOND ORDER LINEAR DIFFERENTIAL EQUATIONS
129
The function is not identically zero, but otherwise , and can be arbitrary functions of . In this book we will only consider the case where , and are all constants with . If the right hand side of a linear differential equation is identically zero, then the equation is said to be homogeneous , otherwise it is nonhomogeneous. In the spring problem analysed above, the equation is homogeneous if there is no driving force. EXAMPLE 8.1 The equation
{
is homogeneous and linear.
8.1.1
The homogeneous case
We begin with a general linear homogeneous second order differential equation of the form (8.4)
%
{ ¤ d ð { d ð ð { r
and the solution is which is of the form . In particular, if , then the solution is of the form . This suggests we try exponential functions of this form as trial solutions in the second order case. Put in equation (8.4) to get
å Ó
{
¦ å Ó ¤ d
¦ å p Ó d ¦ å Ó r å Ó p
{ ¦ å Ó
{ ¦ % ß j ¦ ¥ Ó p
{ ¦
Hence
å
This quadratic equation is called the auxiliary equation of the original differential equation (8.4). Note that powers of in the auxiliary equation correspond to orders of differentiation in the differential equation. In particular, is the zeroth derivative and is replaced by 1. If is to be a solution of the differential equation, then the constant has to satisfy the auxiliary equation. EXAMPLE 8.2 Use the method above to ﬁnd two solutions of the differential equation
{ C d ð { d ð ð { ¦ å Ó {
%
¦
¦

C
C
å
å
d
Ó
Ó
d

C
C
d
+
d
¦

d
å
Ó
d
d
+
d
¦
+
å
Ó
å
Ó
{
Solution. Substitute
into the differential equation. Then
¤
d
d
r
¦
Ó
and, since
is never zero, we have
¦
å
Ó
{

r
¤
B
d
ð
{
d
r
+
¤
{
¦
d
ð
{
k
r
where case. If
r
, and are constants. To get solutions of this equation we generalise from the ﬁrst order , then
ò
õ
ô
C
{
B
ð
ð
The equation
!
is nonhomogeneous and linear.
k
r
¤
r
¤
r
{
7
d
ð
{
7
d
ð
ð
{

+
r
!
130
OSCILLATION PROBLEMS
Let and be any functions. If and are constants, we call a linear combination of and . In the example above, a linear combination of two solutions to the differential equation was also a solution. This is a consequence of a general result known as the principle of superposition . THEOREM 8.1 The principle of superposition Let and be any two solutions of the linear homogeneous equation
{ ¨ { ¨ { { ¤ d ¤
This theorem can be proved by using substitution to check that satisﬁes the differential equation. We will need another result from the theory of differential equations. We shall assume this result without proof. THEOREM 8.2 For any second order linear homogeneous differential equation, we can ﬁnd two solutions which are not constant multiples of each other. If these solutions are and , then every solution has the form for arbitrary constants and . As in the ﬁrst order case (page 103), a solution of a differential equation containing arbitrary constants, with the property that every solution of the equation can be obtained by taking particular values for the constants, is called the general solution . We have shown how to use the auxiliary equation to ﬁnd solutions to a particular linear homogeneous differential equation with constant coefﬁcients. However, an arbitrary linear homogeneous differential equation will have solutions belonging to one of three classes of functions, depending on the nature of the solutions of the auxiliary equation. Case 1: Distinct roots of the auxiliary equation
å Ó { { ¤ £ { ¤ d { £ { ¤ d {
This is the easiest case to deal with. Suppose the roots and are real and different. Then and are both solutions of the differential equation and the general solution is
7 ¦ o å Ó ¤ d ¦ å Ó £ { ¦ o å Ó
¤
where
and
are arbitrary constants.
¦
{
{
{
¤
d
{
where , and are constants with . Then any linear combination also a solution of the differential equation.
£ % k r ¤ r
of
and
is
i
£
§
{
£
7
%
{
¤
¤
and
¥
d
Ó
)
¨
£
so that is a solution. Similarly, for any constants put straightforward to check.
{ £ ¦ ¥ Ó ¤ d ¦ ¥ ¥ Ó Ó £ { § {
is also a solution. More interesting, if we , then we ﬁnd is also a solution. This is
¦
¥
Ó
g
d
¦
¦
¥
¥
Ó
Ó
¥
'
Ó
f
{
and or . Thus both and equation. We can check this by direct substitution. If we put
¦ ¥ { ¦ ¥ ¥ Ó ¦ ¥ { ¨ ¦ ¥ g ¨ f " f
¦
7
Ó
%
¦
Ó
¦ g d E d Ó ) C ¨ ) ¨ r ' d E Ó C 2 ¨ ¦ £
are solutions of the differential , then
SECOND ORDER LINEAR DIFFERENTIAL EQUATIONS
131
EXAMPLE 8.3 Solve the equation The ﬁrst step is to ﬁnd the auxiliary equation. In the above differential equation, we replace , by and by . This gives
2 % g " f d " E ) C 2 E ) C ð 2 % E ) C 2 g " f E ) C ð 2 d E ) C ð £ " £ a £ E U £ % ) U f C f f ð 2 ð ¤ ´ { g 2 ' { 2 2 " ` ð d 2 ð 2
by
E
EXAMPLE 8.4 Solve the equation The auxiliary equation is is
% { ` f ð { r f ð ð {
EXAMPLE 8.5 Solve the initial value problem
% E % C 2 7 " E % C ð 2 7 % f E ) C ð 2 d E ) C ð ð 2
As in Example 8.3, we ﬁnd the general solution to be
¦ § Ó ¤ d ¦
Note that a differential equation has many solutions. These may all be written in one form as the general solution. A differential equation together with initial conditions has just one solution, in which the arbitrary constants in the general solution have been given particular values.
E
¦
°
¥
Ó
f
¦
§
Ó
C
¤
´
f
which we solve to get
and
. The ﬁnal solution is
¤
"
%
¤
¤
'
d
d
£
U
f
"
£
%
)
"
ð
2
Put
and
to get
We now have two equations in
and
:
%
)
%
2
£
¦
§
Ó
¤
'
¤
and we have to use the initial conditions to ﬁnd gives Next, differentiate (8.5) to get
Ó £ ¤ d £ %
and
. In equation (8.5), put
"
f
and its roots are
¦ Ó ¤ d ¦ ¥ Ó
¦
%
§
Ó
¤
{
d
`
d
¦
°
¦
f
°
°
¥
ð
¥
¥
{
Ó
£
Ó
r
f
£
'
r
f
U
f
The roots are as
and
. The solution to the differential equation can now be written down
and 6, so the general solution of
and
. This
)
C
ð
ð
(8.5)
132
OSCILLATION PROBLEMS
EXAMPLE 8.6 Solve the equation This is a third order differential equation, but it is linear, homogeneous and has constant coefﬁcients. The ﬁrst step is still to ﬁnd the auxiliary equation. In the above equation, we replace by , by and by . This gives
§ E ) C ð ð ð 2 % ` d a f § " E ) C 2 E ) C ð 2 % E ) C 2 ` d E ) C ð 2 a f E ) C ð { E g C C d d ' f ¦ E d d f § d Ó ¤ ð ¦ ð ¦ § £ ` § { d f d å r å d Ó Ó ð ð ð ð ð C r r g ¤ f d 2
The roots are now a bit harder to ﬁnd. One way is to try various values for such as and see if we can get lucky. Once we ﬁnd one root, we can do a long division and end up with a quadratic, which we can solve. For the above auxiliary equation, we ﬁnd is a solution. This is done by long division of by . We have
" E ` " f f d C ` E " d f a a f § 7 g ® 7 " ® 7 %
Case 2: Equal roots of the auxiliary equation Suppose the auxiliary equation has a repeated root . Then is one solution and in order to get the general solution we need a second independent solution. We can do this by using the method of variation of parameters introduced in Chapter 6 on page 104. , where we replace the constant in the ﬁrst solution by a function We try a solution . Then
å Ó £ { ¦ § å Ó ¦ § ð å { Ó E ) C { ¦ § £
A rearrangement of this equation yields (8.6)
% ¤ d d % ð ð r d E r g d C ð d r C
.
E
d
g
f
C
r
Secondly, this equation gives
d r 3
%
¤
d
d
r
Firstly,
3
. . Equating coefﬁcients in
r
Now we use the fact that implies two results:
is a repeated root of the auxiliary equation
%
¤
d
E
%
d
ð
{
C
¤
d
d
ð
Substitution in the differential equation
ð g d ð ð C r
and cancellation of the exponential gives
¦
§
å
Ó
¦
§
¥
d
Ó
¦
¥
§
å
d
Ó
¦
Ó
ð
¤
E
2
ð
g
ð
{
"
The roots are written down as
,
and
. The solution to the differential equation can now be
E
'
d
C
E
g
f
C
E
"
f
. This
SECOND ORDER LINEAR DIFFERENTIAL EQUATIONS 133 This is. Thus a second solution to the differential equation is r g f % ¤ d d r . ) ' E ) C { § ) d E ) C ð ð { U d E ) C ð ð ð { % { f ð { d ð ð 5. Once this rule for the second solution has been found. The auxiliary equation is The general solution of the differential equation is § j ¦ ¥ Ó £ { % { d ð { ` d ð ð { Case 3: No real roots of the auxiliary equation satisfy the ) g d g f The auxiliary equation arose by considering exponential functions as trial solutions. ) are not constant multiples of each other. 3. % % E 2 C { d E 2 C ð { E 2 C { U d E 2 C ð ð { ) ò õ ô E ) C 2 d E ) { C ð f 2 ð ) { U d E d ) ð C ð ð ð { ' 2 1. " E ) C 2 ' d E ) C ð 2 ' E 2 d ð ð ð 2 C 7. ) ò õ ô Ó ) ó Î ò Ó How would we ﬁnd the solutions 3 and if they were not known? ò õ ô . there is no prior guarantee that replacing a constant by a function will produce a second solution. % % { { U ' d f ð ð { U { d d ð ð ð ð % % { ' { f d ð ð { { g g d d ð ð ð 11. ﬁnd the particular solution satisfying the initial conditions. { Solve the following differential equations. Otherwise. ) ' d ° ) U E ) C ð ð 4. However. % ð { r d ð ð { % { g d ð { ' d ð ð { 9. The auxiliary equation is An example of this case is the equation which has no real roots. in fact. EXERCISES 8. the general solution since the two solutions (the case . { ` E % C ð { 7 ' E % C { { 7 % ð { ' d ð ð 15. . EXAMPLE 8. 6. it is easily checked that both and differential equation. Historically. In the present case for example.6) enables us to deduce . What is the connection between these and the trigonometric functions? 3 We will address these and other matters in the following sections. Where initial conditions are given. so that for arbitrary constants and . If the equation is nonlinear. 13. a good deal of trial and error was used in discovering methods for solving particular types of differential equations. Since is an assumption made at the outset. 2. However.7 Let . This raises two issues: ¦ Ó ) ó Î ò ¦ Ó % ¦ { g d ð { g f ð ¦ ð { % ' ~ " f 7 ' ~ " f with roots % ¦ § å Ó ) % " ð ð ¤ % ¥ ¦ " § ¦ § å d Ó å § E j Ó ) ¦ ` ¥ ¥ Ó d ) d ¤ ¤ C d ¦ § ¥ å Ó E ) C ¤ { % k % r ¤ " ) ¥ % d ¥ ð ¤ ð r that E ) C Using the results and in equation (8. 8. ) and (the case . we conclude that . ﬁnd the general solution.1 In Exercises 1–8. give a reason. it is easy to apply. 14. state whether the equation is linear or nonlinear. as we have just shown. { 12. it does. { { 10.
{ C B ð { d ð ð { < { < d ð { H B ð ð { 37. 28. but this is historically inaccurate. 7 B ð = d ð ð C { B ð { d ð ð { 31. For reasons which we shall not detail here.2 COMPLEX NUMBERS In order to deal with the case where the auxiliary equation has no real roots. It is often stated that complex numbers were introduced in order to solve quadratic equations such as . 38. 24. 20. { d ð { d ð ð { H B ð ð ð { ð { d ð ð { 7 d ð ð ð { 7 21. { < B ð { d ð ð { C { 7 d ð { C d ð ð { 23. 32. { d ð { < d ð ð { { < d ð { H d ð ð { 35. A formula for solving such equations was published by Gerolamo Cardano in 1545 in Italy and this formula often requires complex numbers as part of the work in reaching the answer.134 OSCILLATION PROBLEMS 8. { d Figure 8. the need for complex numbers was ﬁrst encountered in the sixteenth century in connection with cubic equations. { d ð { B ð ð { B ð ð ð { { d ð { 7 B ð ð { C B ð ð ð { 25. 40. . It was the attempts to understand this aspect of the formula which provided the initial impetus for the study of complex numbers. we have to consider complex numbers. 18. 22. H C d ð C d ð ð 7 B ð B ð ð 33. even if all the roots of the equation are real. 26. 30. { < d ð { d ð ð { { C B ð { B ð ð { 29. { d ð { 7 d ð ð { 7 { 7 B ð { C d ð ð { C 19. Mathematicians were quite happy to accept that such an equation had no real roots and there was no compelling reason to consider any other possibilities.  + ð { % + { % { C d ð { B ð ð { 16. { 7 B ð { d ð ð { { < d ð { B ð ð { 17.2: The complex number  + Im  + Re ü = d ð ü d ð ð ü { d ð { 7 B ð ð { 7 39. { d ð { < d ð ð { C d ð ð ð { ð { B ð ð ð { 27. 34. 36.
then we deﬁne by ± % { ± % ± + ± ± is a (8. If is a complex number. as the case may be) the corresponding real and imaginary parts of the two complex numbers. multiply and divide them. The rules for addition and subtraction are straightforward. Let us now be more precise. Complex numbers are often denoted by the symbols or . we have to be able to add. We can then think of a complex number as a point in the plane.COMPLEX NUMBERS 135 The new idea that occurs with regard to complex numbers is that we go from the usual number line to a number plane. A complex number is an ordered pair of real numbers. then  { % + There is also a simple rule for multiplying complex numbers by real numbers.7) E    H H B + B % C d + %  d C C  B d % % + + + d  H B % C + EXAMPLE 8. If complex number and is any real number. then is called its real part and is denoted Re . We add (or subtract. and (with ) are all complex numbers. The rule for multiplying two complex numbers is more complicated. then . it is called an Argand diagram or complex plane (Figure 8. that is. In order to do anything useful with complex numbers. For example . Complex numbers can be represented by points in the plane and when the plane is used for this purpose. then % { d { % + d + + d h { { g % { % + + { % + { % + + { % % + + + + 7 % + so for example if . but the reasons for it will become clear soon enough. then  { B { % B + B .2). we consider pairs of real numbers. then we deﬁne { d { % { + % { { 7 B B % C + { + % + 7 B % + E  7 § C d < § % <  < § % C 7 + B §  7  < C C § % % + + +  < % 7 + EXAMPLE 8.8 and If % + . subtract.9 and If C % + . Thus if and . If and . Two complex numbers and are equal if Re Re and Im Im . The imaginary part of is and is written Im .
we write . we write for the complex number . Gottfried Leibniz In 1806. but by the end of the second decade of the nineteenth century the idea had become familiar to mathematicians. 8. the complex numbers of the form behave like ordinary real numbers. This was not always the way in which they were discussed.8) B % = % + 7 % d r + % B + r § d § % % = + § < B % C 7 § + § + d % % + % + + % r + % + % < + r % r + % C + % + which we can identify as the number . For example is equivalent to the statement and is equivalent to . Since no real number has this property. Thus in the Argand plane. we can use { d  % + { d { % d + . We identify the complex number with the real number . Firstly. the axis with the point Now for the amazing bit of magic that gives the whole system its remarkable properties. the number more precisely.9) becomes the mysterious looking result B B % + % + . We can write this fact as % + B % + B % B + or. Complex numbers are simply ordered pairs of real numbers with certain rules of arithmetic. They are alson to think of this idea appears to have been Kaspar Wessel. then equation (8. If we square the complex number we get (8. Initially this also went unnoticed. the real numbers lie along corresponding to the real number . The approach to complex numbers that we have used is logically sound.136 OSCILLATION PROBLEMS There are some interesting consequences of these arithmetical rules. In particular. not being. (8. If is any complex number. most an amphibian between being and a Norwegian surveyor. If we write complex numbers in the form . there are severe logical difﬁculties with this approach. in 1797.2. namely . the number . as is commonly done. Thus in the complex number system. For instance . complex numbers were introduced by simply deﬁning to be a number having the property . then we can also ﬁnd the square root of any negative number. The ﬁrst perful ﬂight of God’s spirit.1 The number for the complex number gives us an alternative notation for complex numbers. If. Historically. we can write { % + d % + { % + % + B B B + + · B Once we can ﬁnd Writing { % + This notation is almost always used in preference to the ordered pair notation because it is more convenient for computational purposes. has a square root. It was not until the ordered pair deﬁnition was thought of that comThe imaginary numbers are a wonderplex numbers were put on a ﬁrm logical base. There is nothing mysterious about them.9) Notice that the square root of is not a real number. but his work went unnoticed. They can be represented on diagrams and visualised in a concrete manner. Jean Argand published an account of the graphical representation of complex numbers.
Indeed. 8. It is not clear that this example achieves anything useful. the point axis. This yields { 2 d { {  2 { { B r { r d d B d B ¤ { { r 2 { { 2 all the normal rules of algebra.10 Find all the roots of the equation Solution. if complex numbers had never got beyond this stage.10) came ﬁrst. So for example.3). Historically. that Let . Let the length of this line be . { % + E  7 B 2 C ® ® B + B ¤ r with . if 7 + B 2 2 2 2 B + 2 B d d d + r d C { { { { { 2 { { d d ® d d { 2 + r r B d { 2 d . especially in electrical and electromagnetic applications. We use the usual quadratic formula: B d d .COMPLEX NUMBERS 137 (8. but from a logical point of view it needs to come second. then { 2 d + { 2 d + { 2 d .1 So far. From now on we shall use the form of notation. we have % { d { % { { B + which is just the rule for multiplication we gave in equation (8.10)  If we write this last statement in the ordered pair notation. the result (8. Plot in the complex plane (Figure 8. This is best handled by the notation of this section.2. We join . We will soon show that we can do far more with complex numbers than simply write down solutions of polynomial equations. 1 Sometimes the symbol is used instead of .2 Terminology ~ { 2 d be a given complex number. they would have been abandoned a long time ago. we have not dealt with the matter of division of complex numbers. EXAMPLE 8. We have { 2 d { 2 B + 2 d r + 2 d r This process is called realising the denominator. with the additional rule that and .7). This is to avoid confusion with the use of for electric current. to the origin by a straight line making an angle with the positive direction of the is.
11 If and are any complex numbers. Notice that if ¡ 2 3 . We take to be positive if it is measured counterclockwise from the direction of the positive axis and negative if it is measured clockwise from the direction of the positive axis.138 OSCILLATION PROBLEMS Figure 8. Either of the two expressions or cis is called the polar form of . { f 2 2 E ~ C ó Î ò 2 { E ~ C ò ö ¤ õ ô 2 E g 7 % 3 3 EXAMPLE 8. modulus and argument of a complex number The following terminology is used.3: The conjugate.4. show that ¡ d ¡ d 3 ¡ The complex conjugate of is the complex number plane it is the reﬂection of in the axis. We call the Cartesian form of . we have and . Its value lies in the interval . We write or Ph . This is often abbreviated to cis .   The modulus or amplitude of 3 is the length . ~ ~ { d 2 ¡ { ~ ¢ £ ~ 2 ~ .1). From simple trigonometry (Section 7. When drawn in the complex is real then . Thus . By Pythagoras’ theorem · ~ ~ E E ~ ~ ó Î ó Î ò ò  ¡ ¡  d d ~ ~ ò ò õ õ ô ô C C ¡ 3 3 . We write it as { d 2   The argument or phase of is the angle .
We put and .COMPLEX NUMBERS 139 Next we have E { 2 d 2 C E { 2 d 2 C Finally E { 2 d 2 C E { 2 f 2 C ¡ There is a useful fact about complex roots of polynomial functions which we shall need later. Then . 7 g ~ " g f ~ E ó " Î ò f C d 7 " g · ~ "   ~ ò õ ô 2 f " EXAMPLE 8. This is that complex roots always occur in conjugate pairs if the coefﬁcients in the polynomial are real. and are real. ÿ ÿ ' f þ ó Î ò 2 d ÿ ' 2 f ' þ ò U õ ô þ f ! U { 2 ç d § 2 f æ EXAMPLE 8. where . so that and . % ¤ d ¡ d ¡ r We also have E U ~ a C g U ~ a so and cis . Consequently ¤ r ¤ d d 7 ¡ % r ¤ d d r % ¤ d d r that is. Suppose for example that is a zero of the polynomial .12 The Cartesian form of cis ! ¡ Thus is also a zero of the polynomial.13 Find the polar form of Solution. is given by i { 2 f 2 ¡ { 2 E E f E E { { 2 { { E E 2 2 2 2 { { d d f d ¡ d d E { { 2 2 C C { { { 2 2 2 C C C C d d 2 2 2 2 f E E f d f d { { 2 2 2 C ¡ { 2 2 { { { E 2 f d d  d d { { d { d 2 2 2 f f ¡  C C 2 2 2 f 2 2 2 ¡ 2 ¡ C 2 2 d { 2 d 2 Solution. We have . Then ~ .
Evaluate each of the real or complex numbers in Exercises 1–21. Re  §  2. 21. then we can write and 7 ' 7 g 7 " ± { 2 d 2 { ÷ Ñ W 2 d 7 2 ' 7 g ÷ Ñ 7 W " ± where is a complex number for each then we can write Ñ W If we put for % " d 2 d ° % ÷ " f ° ÷ % " f § 31. { ÷ Ñ X 2 ÷ Ñ X 2 d r { ÷ Ñ W ÷ Ñ ÷ X Ñ X We say that the complex series converges. 26. 5. If and 2 ÷ Ñ X r converges if each of the real sums . 11. . r d § g f § E § d C 13.3 COMPLEX SERIES Let us take an inﬁnite sum of the form 7 Ñ W for the sum of the complex series. d 1. 6. 24.     d    d  7. Im { 2 d 2 2 g f " § 2 ' f " 2 d g . and . ú ú § ú ú § § d 19. 12.140 OSCILLATION PROBLEMS EXERCISES 8. express the given complex numbers in polar form.2 Let . E § C §       10. 32. 14. 2 U d U ' 2 d " 2 g f ' g 22. 8. 9. 2 f " f 2 ' f 2 g d g f 25. Solve the equations: % " d § 2 % " d 2 g d 2 ' % U d 2 d 2 28. E § C 4. Im f 3. 23. 2 8. " Ê É 16. 27. 20. Where a number is complex. 2 29. 15. Re 17. 2 30. 33. ú ú ú ú In Exercises 22–27. express it in the form . d 18.
12) { { Euler’s equation gives rise to yet another remarkable identity.3.COMPLEX SERIES 141 8. If we put " Ó we get { ó Î ò Similarly. This is a surprising and remarkable result. we can show that the series for converges for any complex number . if we ﬁnd the difference of equations (8.12) and solve for ¥ Ó Ó ¥ { Ó ò If we add equations (8.1 The complex exponential 2 d g f d d d f õ d f ¯ ¯ § d § ô E { ¯ The ratio test applies to complex series as well as real series. we ﬁnd . this series is hard to sum as it stands.11) { for any real number . For a given complex number .11) was discovered by Leonhard Euler in the eighteenth century and is known as Euler’s formula . From Euler’s formula we have " 7 E f { f C ó Î ò { f C ò õ ô ¥ Ó { Ó so that (8. we ﬁnd ¦ ¦ ¦ Ê d ¦ ¦ ¦ ¯ ¦ d ¦ ¯ ³ ¦ r { d d ¦ d ¦ ¯ ¤ ¦ ¯ § ' { d d ó f ¦ Î ó ¦ ¯ ¦ Î ò ¦ { ò ¦ U ¦ É g ° d ' { ò U { ò õ ¯ Ê õ d ô ¦ ' d ô ¦ ¯ ¦ ¯ § g { { ' { ò d d ó d Î õ ¯ ¥ ° ò ô Ó f ¯ { Ó U { g ¯ ó d Î g { d ò " ¯ f d g { d { { § f { Ó ò " d d õ É " " ô § Ó In a similar way. Then ¤ E { C ° E { C § E { C E { C h g { § Ó { We have shown that (8.12) and solve for Ó . All very strange.11) and (8. Using this test. Equation (8. but if we rewrite it we can make things simpler. Yet here we have a connection between these three functions and the square root of . Consider ﬁrst the case where with .11) and (8. The notation stands for Euler’s number. For real functions there seems to be no connection between the exponential function and the sine and cosine functions. deﬁne for all complex numbers by letting ¦ ¦ ¦ d ¯ ¤ 2 d ¦ ¦ ¦ d ' 2 d ¯ g 2 d 2 d " Ó Ó We have previously deﬁned for all real numbers by letting Ó .
In particular.142 OSCILLATION PROBLEMS Here we have four of the most important numbers in mathematics connected by one very simple equation. so that ¨ Ú § Ó r ¨ Ó § Ó Ú Ó § Ó Ú ß Ó Ó ß Ó EXAMPLE 8. E { ó Î ò d { ò õ ô C E { ó Î ò d { ò õ ô C o Ó Ó so that E ~ g ó Î ò d ~ g ò õ ô C E ~ ó Î ò d ~ ò õ ô C More generally.14 cis . then If E ~ ó Î ò d ~ ò õ ô C ~ So to multiply two complex numbers.13) . We showed earlier that for any real numbers and . then we see that i E { ò õ ô { ó Î ò d { ò õ ô { ò U õ o ô © C Ú E © ' E d T { Ó { d d ó Î " ò { C C { g ó Ó Î ò ò õ ô d f E { { ò U õ d o ô Ú o { { © C ò ò T Ó õ õ ô ô Ó ~ © Ó If and . we multiply their moduli and add their arguments. we have identity holds for complex exponents and . E ~ ó Î ò d ~ ò õ ô C ¤ E ~ ó Î ò d ~ ò õ ô C © Ó E © Ó C { { If we take in this result. we can use the equation © ¤ Ó ¤ E © Ó C to show that This last result is known as de Moivre’s Theorem.15 . then E { ó Î ò d { ò õ © ô C Ó Ó Ó Ó (8. EXAMPLE 8. It can be shown that the same . then If E ' ~ ó Î ò d E ' ~ ò õ ô C Ó § j Ú Ó ' ~ d g As a consequence of Euler’s formula (equation (8.11)) multiplication of complex exponential functions is intimately connected with addition formulas for trigonometric functions.
show that and for i ~ § ó ~ Î ò § ó § Î 2 ò 2 d ~ f ~ ó Î ò ó Î ~ ò ò ~ õ ~ ò ô § õ ô ó 2 Î ' ' ò ô ô ~ . Then the general solution to the differential equation (8. EXERCISES 8. We are assuming the in the case where the auxiliary equation coefﬁcients . and are real so that the roots occur in conjugate pairs. Prove that the solutions of g 7 " are given by E ~ ó Î ò 2 d ~ ò d " ó Î ò 2 d ò õ ô 2. Let be the roots of the auxiliary equation. If ~ j 1. . We have C ~ ' ó Î ò 2 d ~ ' ò õ ô If we equate the real part of the left hand side to the real part of the right hand side we get 7 ~ ó Î ò ô while equating the imaginary parts of both sides gives ~ § ó Î ò Notice the way in which complex numbers have been used as tools in obtaining results which are purely real.4 COMPLEX ROOTS OF THE AUXILIARY EQUATION We can now return to the differential equation (8.14) ( % has complex roots.14) can be written in complex form as 2 ® 1 ¦ h ) ¥ ¤ r E r ~ d C ò õ ô g E r ~ C ò õ ô ° 4.COMPLEX ROOTS OF THE AUXILIARY EQUATION 143 EXAMPLE 8.16 Use de Moivre’s Theorem to show that ~ ó Î ò ~ 7 g ò õ ô ' f ~ % d ~ õ f d g f f ò ô 2 C ~ ~ Ó ~ ~ õ E ° ô { ¤ ó ó ó ó ò ~ ¦ ¤ Î Î Î Î ' õ ò ò ò ò ò d ô d õ d ~ ~ ~ ~ f § ô ¦ ! ð E % ~ U g { ò ò ò ò h ~ § ) § õ õ õ õ ò ó ò f ô ô d Ú ô ô Î d õ õ d g 2 2 ò " " ô ' ' 2 ô ð ' ' C 2 ð Ó { ò d d d ¥ d r õ r ~ ' ~ ~ ô ~ ~ ~ ~ ' ' ' ò § § g ' ò ò ò ò ó ó õ ' { Î Î õ õ õ õ ô d f ò ò ô Solution.3  § ) Ó  § ) Ó ` 3. evaluate . Use de Moivre’s theorem to show that " f 2 ò õ ô ! " 2 ` ò õ ô 8. If .
1. In the above equation. Then the general solution of this differential equation is ) ( ó Î ò ¦ g Ó ¤ d ) ( ò õ ô ¦ g Ó £ { ¤ r Once again.3 are the roots of the auxiliary equation of the homogeneous linear equation Suppose 7 % { ¤ d ð { d ð ð { r ( 2 ® 1 where . since we want real solutions.144 OSCILLATION PROBLEMS where and are arbitrary constants.17 Solve the equation % E ) C 2 ' " d E ) C ð 2 ` d E ) C ð ð 2 3 2 g ® " f % r d E ) g g ó d Î ò The equation has auxiliary equation so the general solution of the differential equation is ¤ d ) g ò õ ô £ C ¦ ¥ Ó { % { r d { g d ð ð { 3 with roots 2 ® " f % E ) g ó d Let . However. E ¦ f C 2 ¤ ¦ d are real if ¥ ç E ) ( ó Î ò 2 ç ) f ( ) ó Î ( ò ò E õ ô £ ¦ C f ¦ 7 d ¥ ) C ( E 2 ) ó ¦ ( Î h E d ò ) ¦ ó ) ¦ h Î ¥ ) ( g ò Ó Ó ¥ ¦ ò 2 g Ó õ ¤ Ó ô d ¦ E ) d ¦ ( d ) ¦ ò d ( ¦ õ h ¦ d ò ) ô h õ C ) Ó ô ¥ Ó ¥ C ¥ ¦ ¥ ¦ æ æ C g g ¦ ¦ ¦ Ó Ó g g g ¥ Ó Ó Ó £ ¥ ¥ £ and are suitably . We have come to an important result.) This gives E ) C ð ð 2 % ' " d ` d " E ) C 2 E ) C ð 2 ¦ ¤ where and . we can rewrite it in real form using the results of the previous section. we replace by by and by . we have used complex numbers as a tool to obtain a real result.1 in the text justiﬁes this procedure. EXAMPLE 8. (Section 8. . This form of the solution is not particularly useful for our needs. and are constants. The auxiliary equation is Hence the general solution of the differential equation is Î ò g ¤ d d ) ò õ ô £ C ¦ ¥ Ó { % { g d ð { g d ð ð { 3 E ) g ó Î ò ¤ d ) g ò õ ô £ C ¦ § ¥ Ó 2 2 g ® ' f The roots are as . . . We have ¦ h ) ¥ g Ó ¦ d ¦ h ) Ú g Ó ¥ { ¦ THEOREM 8. The constants and chosen complex numbers. The solution to the differential equation can now be written down and this has roots The ﬁrst step is to ﬁnd the auxiliary equation.
The exact nature of the solutions of this equation will depend on and and we will need to break down the problem into a number of ) and the % E g % C f ð { % E 7 % " C E ð " f { C ð { 7 E g % 7 C " { E % 7 C E % % % % { % C { 7 { { { { { % % r U g 7 % d d % d d f ð ð { 6. it is convenient to write this equation in the form ) C ò õ ô 6 { % ' Case 1: Simple harmonic motion ( % ' where the driving force the values of the coefﬁcients cases. ﬁnd the general solution. 5. ﬁnd the particular solution satisfying the initial conditions.5 SIMPLE HARMONIC MOTION AND DAMPING We return to the problem of a particle oscillating on a spring. 10. 3.SIMPLE HARMONIC MOTION AND DAMPING 145 EXERCISES 8. 15. This is the case equation becomes { d ð ð { Motion governed by this equation is called simple harmonic motion (SHM). we have only developed the techniques to solve the homogeneous case (8. 16. 11.4 Solve the following differential equations. 14. 8. The general oscillation problem requires the solution to the differential equation { d ð { ' g d ð ð { At this stage. E ) C û . Otherwise. The auxiliary equation is with roots and so the solution is ¤ d ) ò õ ô £ { ® % In order to analyse the motion of the mass. . 12. % { d ð % ' % { U % % % f ð { { { a ' { ` " r ' % % d d d d ð ð ð ð ð { { { { { E U { ) 7 g U U { U ) C % f f E f f f d d û ó ð ð Î ð ð ð ð ð ð ð ð ð ð ð ð ð ð ò % { { { { { { { f { % { r d ð { g d ð ð ð { ð { { ð g g ` { { { { " g g { f d d f d d d d d ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð d { { g ' { { { { { { { { 1. 9.15) occurs when there is no friction.15) d ð { ' g d ð ð { The simplest case of equation (8. 13. 17. 7. Where initial conditions are given. % 8. 4. 2.
care must be taken to choose the correct quadrant. To do this. This distance is called the amplitude of the motion.4: Simple harmonic motion displacement of the body from the equilibrium position. The constants and have direct physical interpretations.16) g f . The constant gives the maximum U ~ g 6 6 ó Î ò 6 ò õ ô Figure 8. we write (8. The phase angle determines the initial position of the body. The constant is related to the . The constant is the phase angle . A frequency of 1 Hz corresponds to 1 cycle per second.18 The motion of a mass on the end of a spring is given by the initial value problem % E % C ð { 7 g E % C { 7 % { d ð ð { ~ ' % { 2 The unit of frequency is the Hertz (Hz). Á g E 7 E U ~ ó Î ò f ) ) C ò ó õ Î ô ò g d E { ò õ f ô ) ) C ò ò õ õ ô ô C ¨ 6 6 " " g g f f ) ó Î ò ¤ f d ) ò õ % ô w £ 6 for certain constants and . . Figure 8.16). For instance. if the body starts at the maximum displacement below the equilibrium the body starts at the maximum displacement above the equilibrium position.4 shows the solution for the case . while if since is positive when the body is below the equilibrium position. position.146 OSCILLATION PROBLEMS so that ó Î ò 6 ¤ 7 ò õ ô 6 £ Hence £ ¤ ó . frequency2 in cycles per second by EXAMPLE 8. 7 ¤ d £ · 6 In calculating . This can be done by checking the signs of and in equation (8.
(We cannot have . u 9 u % Case 2: Weak damping ( ) are E U f % d 9 g d The roots of the auxiliary equation Since . are .) The equation shows that 6 6 7 % ) g ò õ ô g ó F { ò õ ô 6 g ó Î ò 6 % % ) % w g 6 ~ g { ò õ ô E f ) C ó Î ò 6 f ð { the same solution as before. According to the discussion above. since % E % C ð { g ) ó E Î % ò C { ¤ d ) ò õ ô £ { £ . We now have and put and . Hence the solution % ) % ð { % ò g õ ô 9 ) U ò f õ · ô 9 g ® · 9 g ® { f 9 C % f % ) % ð { If we put and . we can write the solution in the form 7 E f ) C ò õ ô 6 { so that We put and in the ﬁrst equation to get second equation to get .19 As an example we take the equation The roots of the auxiliary equation is j ¦ ¥ Ó { with initial conditions .5. so . .SIMPLE HARMONIC MOTION AND DAMPING 147 The physical interpretation of this initial value problem is that the mass starts from rest ( ) at a distance of 2 units below its equilibrium position ( ). The solution is ) 7 ) ò õ ó ô Î ò ¤ ¤ d d ) ) ó ò Î g ò õ ô g g f ð { { % ) and g in this equation to get . Hence the roots of the auxiliary equation form a complex conjugate pair of the form and the solution is ( 2 ) ® ( 1 ó Î ò ¦ g Ó ¤ d ) ( ò õ ô ¦ g Ó £ f 9 { · u 9 u % and using the initial conditions we ﬁnd ) " " ' ó Î ò " d ) " " ' ò õ ô j ¦ ¥ Ó g { This solution is plotted in Figure 8. Thus we end up with in the . the number is nonzero and pure imaginary. Hence . we ﬁnd . the solution is Put { so that ¤ Alternatively. % E % C ð { g E % C { g ~ E 2 " " ' ) " ® " g " g f ' C ó Î ò " " ¤ % d ' ) % " " { g r g r ' g g ò d õ d ð ô { £ d d ð ð { EXAMPLE 8.
the number is nonzero and real. % w w 9 Case 3: Strong damping ( ) are E U f % d 9 g d The roots of the auxiliary equation EXAMPLE 8.6.5: Weak damping The behaviour shown in this example is typical. ¦ ° ¥ Ó § f ¦ ¥ Ó µ § { U f 7 " f ¦ ° ¥ Ó % ¤ d ¦ U ¥ Ó d £ r { d The roots of the auxiliary equation solution is are found to be .20 Consider the initial value problem % E % C ð { 7 g E % C { 7 % { U d ð { r d ð ð { Using the initial conditions we get The solution is plotted in Figure 8. The body still oscillates. Hence the roots of the auxiliary equation are real and different and the solution is Á g 9 U f · 9 ® · 9 g ® f 9 C f " " g % ¨ f .148 OSCILLATION PROBLEMS Figure 8. but the oscillations die away exponentially. so the general ¦ o å Ó ¤ d ¦ å Ó £ { f 9 · w If and 9 . The frequency of the oscillations is changed by the damping.
As in the case of strong damping.21 We take the same equation as in the above example. but change the initial conditions to . The particle does not oscillate. but if the damping were any weaker it would.SIMPLE HARMONIC MOTION AND DAMPING 149 Figure 8. In critical damping the roots of the auxiliary equation are equal and we get solutions of the type E ) ¤ d £ C ¦ å ¥ Ó { ' E % C Á Á a a ` ` r r U U § ' ' § g g " " " " g g % ' % ¨ . It can be shown that the graph of the solution to an overdamped problem can have at most one turning point. ¨ . % w 9 Case 4: Critical damping ( ) This is the transitional case between weak damping and strong damping. Figure 8.6: Strong damping EXAMPLE 8.7: Strong damping Both these examples show typical behaviour for strong damping. the graph of the solution to a critically damped problem can have at most one turning point. The motion is therefore nonoscillatory.7. You should verify for yourself that the solution is now { ¦ ° ¥ Ó ´ d ¦ ¥ Ó { % " f E % C ð { This is plotted in Figure 8.
22 If then the auxiliary equation is which becomes when we apply the initial conditions.8. Its graph is plotted in Figure 8.9.23 Consider the initial value problem U " " You are asked to show in Exercises 8.9: Critical damping Á % " Á g a f and it has a repeated root of 7 ¦ ¥ Ó ) ¤ d ¦ % % a ! ` E E % % C C ð ð a r { { 7 7 g g ` U E E % C % C { { r ' 7 ¥ g 7 Ó % % £ U g U { d U { { U d ð U d " ' { ð d U { d d ð ð ð ð g { { " g % " ¨ " % f . ¨ ¦ ¥ Ó ) U d ¦ ¥ Ó U { Figure 8.150 OSCILLATION PROBLEMS EXAMPLE 8. i ) Figure 8. the graph crosses the axis and there is a minimum turning point. The solution is .8: Critical damping EXAMPLE 8. In this case.5 that the solution to this problem is j ¦ ¥ Ó E ) ' f " C " E ) C { The graph of the solution is plotted in Figure 8.
and the equation is is given by .25 A 2 kg mass stretches a spring 2 cm. we have Newtons/metre. The mass–spring system is allowed to oscillate in a medium that imparts a viscous force of 8 N when the speed of the mass is 4 cm/sec. Assuming the resisting force is proportional to the velocity with proportionality constant .14) in the text: 7 % { ± d ð { ' g d ð ð { ~ U % ' l ' EXAMPLE 8. the spring constant 7 2 ± 2 ' is given by ± ± ! ~ U § g From . The general equation of motion is equation (8. We ﬁnd N/m. Assume that the motion is undamped and simple harmonic with a frequency Hz. so that . As the damping gets stronger ( ).24 A 5 kg mass is attached to a spring. We have (on page 147) problem for the equation of motion is { 7 % { % U d ð { % % " d ð ð { % % g . Determine the initial value problem for the equation of motion. The spring constant % { d ð ð { ' ~ ± % where so ' and is a proportionality constant for the damping. g ~ % g ' . EXAMPLE 8.SIMPLE HARMONIC MOTION AND DAMPING 151 As the damping gets weaker ( ) we get closer to simple harmonic motion. In this case there is no damping. The initial value % ! g % % ~ g ± U % % § ! % U ! ~ where m/sec is the acceleration due to gravity. the particle takes longer to return to equilibrium. The mass is set in motion from its equilibrium position with an initial velocity of 8 cm/sec. Mathematica is an ideal tool to experiment with the various cases. we have . Since . What is the value of the spring constant ? Solution. we have ± ± so we need to ﬁnd and . If a mass stretches a spring by a length . The general equation of motion is 7 % { d ð { ' g d ð ð { % % " ! ' % g % E % C ð { 7 ~ % E % C ' g Hence N sec/m. Solution. Inﬁnite damping corresponds to clamping the particle at its initial position. The mass is pulled down from the equilibrium position and allowed to oscillate.
The motion for an undamped spring–mass system is given by % E % C ð 2 2 7. The equation of motion of a particle moving in a straight line is For what values of the constant is the motion oscillatory? 6. The equation of motion of a particle moving in a straight line is ± For what values of the constant is the motion oscillatory? 5. the barrel and spring act like an undamped mass– spring system. what is the value of the spring constant ? 8. A 20 kg mass is attached to a spring. A damper is then engaged and the barrel returns to its initial position without overshoot. A mass of 100 g stretches a spring 5 cm. U .5 m before stopping? (b) Assume that in the return phase. the barrel and spring act like a critically damped mass– spring system. If the frequency of undamped simple harmonic motion is Hertz. Show that the solution to the initial value problem a f E % C ð { 7 " E % C { 7 % { " d ð U ' 7 j E % ¦ " % % C ¥ ' g Ó ± 2 E E 2 ) 2 % ' U 7 C d % f 2 ð d 2 ð " C g 2 7 E " g r ) % d C ð d ð 2 ð ` 2 ð E E " ) ) 2 C C d { 2 U E r ) C ð ) d 2 { ' E ) g C d ð ð ð d ð { is 2. A gun barrel weighs 700 kg and the initial velocity of recoil after ﬁring is 25 m/sec.152 OSCILLATION PROBLEMS EXERCISES 8. What should the value of the damping constant be? 4. . (a) Assume that in the initial phase of recoil. ﬁnd the position of the mass at any subsequent time . If air resistance is neglected. The motion for a damped spring–mass system is given by % E % C ð 2 7 " E ) C ð ð 2 ~ g r 7 U 7 Find the equation of motion and sketch its graph for ` 7 r 7 Find the equation of motion and sketch its graph for . What should be the value of the spring constant so that the barrel travels 1. The barrel recoils against a spring until it stops.5 1. The mass is set in motion from its equilibrium position with a downward velocity of 10 cm/sec. When does the mass ﬁrst return to its equilibrium position? 3.
When the system is in equilibrium.17). g { d { { 3 Lastly we add and . We shall simply state these results. is called the comple . The mass is and the system now oscillates with a frequency 12. We shall only consider the case of periodic driving forces such as or . The equation to be solved is (8. Why are exact simple harmonic motions rare? 13. The mass is then pulled down a further 10 cm and given an upward velocity of 10 cm/sec. . If initial conditions are given they should be used in this general solution to ﬁnd the corresponding values of the constants. Methods of ﬁnding the complementary function have been treated in earlier sections but we still need to develop methods for determining . When does the mass ﬁrst reach its lowest point after being set in motion? 11. A mass of 5 kg is attached to a spring hanging from a ceiling. { This equation is called the complementary equation and its solution mentary function . when suspended from a spring. Calculate the ratio . Give some examples of natural motions that are approximately simple harmonic. 10.6 FORCED OSCILLATIONS In this section we consider the case of an oscillatory motion when there is a nonzero driving force. Give some examples of damping devices in common use. It can be shown that the general solution of equation (8. Hence solve the equation. This is called a particular integral or a particular solution and we denote it by .17) there are results about differential equations which tell us that we need to do three things. EXAMPLE 8.17) is a result we shall assume without proof. The motion for a damped spring–mass system is given by % E % C ð 2 7 " E % C 2 7 % E ) C 2 d E ) C ð ` ) ) 7 r ó ó Î 7 Î g ò 7 % ò U { E ) r C { { û U E ) d C ð { d ð û ð { { ' g d g { ð { d ' ð ð g { d ð ð { 2 % " g d { g E { ) C ð g ð ) ó Î ò 2 Find the equation of motion and sketch its graph for .FORCED OSCILLATIONS 153 9. There are no arbitrary constants in . the spring is stretched 50 cm beyond its natural length .26 Show that § g { { d { is a solution of the equation .17) where is not identically zero although it may be zero for particular values of . A mass oscillates with a frequency of removed and replaced with a second mass . To solve an equation such as (8. 3 There will be two arbitrary constants in the complementary function and these will appear in the general solution . Determine the equation of motion for the mass. We ﬁrst ﬁnd the general solution of the homogeneous equation 3 ) ò ) õ ô r E ) C û E ) C û Next we ﬁnd any one solution of the full equation (8. 8.
We would expect that would have to involve a combination of sine and/or cosine functions. Substitute into the equation . Thus The required general solution is ) ó Î ò § d ) g ó Î ò ¤ d ) g ò õ ô £ g { d { { ) g ó Î ò ¤ d ) g ò õ ô £ { % g { { U d ð ð { 8. You should try this and convince yourself that it does not will produce both and work. which is the solution of the complementary equation . The reason it doesn’t work is that successive differentiations of . This gives £ ) ò õ ô .18).1 The method of undetermined coefﬁcients Consider the problem of ﬁnding a particular integral for an equation of the form ) ó Î ò £ { § r d ð { r d ð ð { r We have to ﬁnd a function which can be substituted into the left hand side to yield the right hand side. and into equation (8.18) ) ó Î ò g { U f ð { ' f ð ð { g { g { g { A natural guess for would be . Next we ﬁnd the complementary function. The constants tion (8.6. substituting polynomials or exponentials in the left hand side would not yield a trigonometric function. EXAMPLE 8. so we try ¤ instead. We have and have to be determined by substitution into the differential equa¤ £ ) ó Î ò ´ ³ f ) ò õ ô ´ § f g { ´ ³ f ¤ ´ § from which and . Hence a particular solution is 7 g % ¤ ¤ r ' f f £ £ r ' f f ) ò õ ô ) ó Î ò Equating coefﬁcients of and on both sides of the equation gives ) ó Î ò g ) ò õ ô E £ U f ¤ ' f £ ð f ð g C { d ) ó ð Î g ò { E g ¤ { We substitute these expressions for U f £ ' d ¤ f C . What is a reasonable guess for ? Clearly.154 OSCILLATION PROBLEMS so that is indeed a particular solution.18) and collect terms to get ) ó Î ò ) ) ó ó Î Î ò ò § ° d ) ) ó ó Î ) ) Î ò 7 ) ò ò ó ) Î õ ó § ò Î ô ó Î ò f ò { ¤ U ¤ f d d d ) ç ) ð ) ) ð ò ó ò { õ Î ó Î õ ô ò ò ô £ § £ æ f f U d g ð ð ) g ð g { ç { { ó ) Î ò ó Î ò £ § æ ) ó ) Î ¨ ò ¨ § £ g { { Solution.27 Find a particular solution of the equation (8.
ó Î If and have the form If and have the form ò ) ò õ ô ò ) ò õ ô 3 do not occur in the complementary function. Some other cases are dealt with in some of the later examples and exercises. equation. we cannot expect .26.FORCED OSCILLATIONS 155 EXAMPLE 8. As we found in Example 8. then the particular integral will ) ó Î ò ¤ d ) ò õ ô £ do occur in the complementary function. We can see what the problem is if we examine the complementary function for the differential .28 Find a particular integral for the equation ) g ó Î ò { U d ð ð g { ) ó Î { Following the method in the previous example. We will give you the rules for ﬁnding a particular integral for equations of the form 7 ) ó Î ò ¤ d ) ò õ ô ¤ { § r d ð ð { r where one (but not both) of the constants or may be zero. Since . so . the complementary function is This is the same function we tried above for . This procedure is known as the method of undetermined coefﬁcients. It turns out that in this case. the correct form of the particular it to also satisfy integral is Taking this as the particular integral we use the product rule to obtain ) g ò õ ô £ U d ) g ó Î ò ¤ U f ) g ó Î ò ) ò õ ô ) ¤ U f ) g ò % õ ô ¤ ) g d ó Î ) ò g ó Î ò % £ { { U d ð ð { ) g ò g õ { ô ) ¤ d { ) g ó Î g ò { ) g ó Î ò ) g d % ð ð { ) The two forms of particular integral illustrated in these examples will sufﬁce for most of our purposes. then the particular integral will ) ó Î ò ) ¤ d ) ò õ ô ) £ i ) g ò õ ô ) ° ¤ f ¤ % so that and Hence the particular integral is 7 ) g ó Î ò ) ) g g ò õ ó Î ô ò £ U { d ¤ U f g { Putting these results for and into gives £ 3 . since we cannot have for all . we conﬁdently try ) g ò õ ô ¤ d ) g ó Î ò £ { U £ r ) ) ) U g g g ¤ d £ d g { ò ð ó f ð { ð Î õ { ) ò { ô g g g { Substitution into the differential equation gives 7 ) g ó Î ò ) g ò õ ô ¤ U d ) g ó Î ò £ U d ¤ U f ) g ó Î ò £ U ð ° ó ð g Î { { ò f ð ) ð g U { f which simpliﬁes to so that something has gone wrong.
that is this driving force that starts the motion and then maintains it. % E % C ð { % E % C { d ) g ó Î ò " § f ) If we apply the initial conditions we ﬁnd problem is g ò õ ô { and § ) ¥ ó Î ò § g { ¥ g § % Thus. The complementary equation is % { U d ð ð { and its solution is Since ò is not in the complementary function. ¤ £ 3 EXAMPLE 8.26.29 Solve the initial value problem " f E % C ð i ) ó Î ò r ) § ó Î f ò § { 7 " E % C { 7 ) ó Î ò { U d ð ð { Solution. We will keep particular examples as simple as possible. We will keep our equations consistent in that they will all start at equilibrium and . and and a particular integral is ) ) ) g ó Î ó ó ò Î Î ò ò ¤ ) ¥ f ó Î d ) ò ) ò g ¤ õ ò ô ' õ £ ô d ) f ò ¥ õ ô g ð g ð g £ { { ' ó Î £ . we take ) ó Î ò ¤ d ) ò õ ô £ { ) Hence ) ò õ ô ¤ d ) ó Î ò £ f ð { Substitution into the differential equation gives 7 ) ó Î ò ) ó Î ò ¤ U d ) ò õ ô £ U d ) ó Î ò ¤ f ) ò õ ô £ f so that ¤ This is the same as the particular integral we assumed in Example 8. The general solution is ) ó Î ò d ) g ó Î ò ¥ d ) g ò õ ô ¥ { d { We will now apply some of the principles above to a variety of oscillation problems with an oscillatory driving force. so the solution of the initial value . The driving force will be and it is with initial velocity zero. but with problems of this type hand calculations can quickly become long and tedious.156 OSCILLATION PROBLEMS Once we have chosen the form of the particular integral. A package such as Mathematica can prove very useful. we substitute it into the differential equation and solve for and .
29. where the ratio of the frequencies is irrational. Here the natural frequency of the system is while the frequency of the driving force is . for example. namely ) ó Î ò { U d ð ð { The general solution is Using the initial conditions . Since one is a rational multiple of the other.10: A periodic forced motion one with period and the other with period .30 We ﬁrst consider the case . This is the situation we considered in Example 8. This will not be the case if we had . This reduces to the initial value (8.6. EXAMPLE 8.10. Figure 8. so .20).31 Next.19) % E % C ð ~ " f % Á E % C ð { ) U % ó Î 7 ò % § d E ) % C g { ó ~ Î g ò " % 7 ) g g f E ó % Î C ò ð { { { 7 % E % C { 7 ) ó Î ò r { d ð { ' g d ð ¨ { r r % d E ð % % d % ð C ð k { f ð { { ð { Rather than considering the general case (8. EXAMPLE 8. The solution is the sum of two oscillating functions. the case problem 7 % E % C { 7 ) ó Î ò r { .2 The case of no friction The initial value problem we are interested in is (8. An example of this case is the equation { U E g C % E % ) ¥ % C ó ð Î { ò ' § and we begin by assuming there is no friction.20) d ) g ó ¥ Î ò ¥ d ) g ò õ ô ¥ ) ó Î ò { U d ð ð { . consider the case . we will consider particular illustrative examples. % gives and The graph of the solution is shown in Figure 8. that is. the motion is periodic.FORCED OSCILLATIONS 157 8.
In a real problem such unbounded oscillations cannot occur—either there is sufﬁcient friction to prevent resonance occurring or the system fails because the oscillations become too large for it to bear. but the second oscillation has amplitude increasing with .11).11: Resonance Substituting into the differential equation gives ô ¤ U d ) g ó Î £ U f so that . a phenomenon known as resonance (Figure 8.158 OSCILLATION PROBLEMS The complementary function is the same as Example 8.30. and the general soution is ò ¥ d ) g ò õ ¥ { As in Example 8. They both have the same frequency.30. Since the right hand side of the differential equation appears in the complementary function. f ) Using the initial conditions gives . this solution is the sum of two oscillations. the particular integral has the form ) g ó Î ò ) ¤ d ) g ò õ ô ) £ g { ) g ó Î ò ¥ d ) g ò õ ô ¥ % f ò ô g ) { { Hence ) g ò õ ô ) ¤ ) g ó Î ò ) ¤ ) U g 7 f ) ò g õ ) g ô g ó Á Î ) ò ) d ò ° õ g ) ô ò g f õ ¤ ) ó ô ) Î U g ) g ) ò ò U ° ó d õ Î ¤ d ) g ó Î ò ) £ f ) g ò õ ô £ ð { % ¤ ° f Figure 8. Î ò and the particular solution ) g ò µ g õ ô ó ) £ µ ¥ U g g ó Î ¥ ò £ U f ¨ " " g g ' ' % ð g ð f f f g { £ . In this case the oscillations increase without bound.
This situation gives rise to beats (Figure 8. Another example is the loud squeals from ampliﬁers which can occur as a result of temporary resonance in the electrical circuits driving the system.12. It is difﬁcult to ﬁnd an easy hand calculation which demonstrates ¨ Figure 8.05 Hz with a slowly varying sinusoidal amplitude of frequency 0. this time of slightly different frequencies. The nature of the solution can be made clearer if we rewrite it using the trigonometric identity ¤ ó Î ò £ ó Î ò g f E ¤ f £ C ò õ ô f E ¤ d £ C ò õ ô If we put ) " g g d g £ and 7 ) " g g f g ¤ then The term has a period of while the term has a period of . We take % E % C ð { 7 % E % C { 7 ) " g ò õ ô { U d ð ð { A calculation similar to the ones above shows that ) " g ò õ ô U U g f ) g ò õ ô U U g { Again this is made up of two oscillations. To guard against resonance. EXAMPLE 8. but it is an ideal situation for a system such as Mathematica.12: Beats beats. it is important to know the natural frequencies of the system and the driving force. This leads to a different type of phenomenon. which occurs when Finally.05 Hz. The graph is plotted in Figure 8.12). Thus the term varies slowly compared with the term and the motion is a rapid oscillation with frequency 2. we consider the case the natural frequency of the system and the driving force are only slightly different. r % g ~ g ) ) r r % % g g ó Î ó Î ò ò r g " r % % ~ g ) r % % ó Î ò ) r % % ó Î ò ' ) r % g ó Î ò ) r % % ó Î ò ! ! U f { Á % U % g g g U % U f f .FORCED OSCILLATIONS 159 The Tacoma Narrows Bridge mentioned in Chapter 1 is such a case.32 .
To see the general structure of the solutions we ﬁrst analyse a case where damping is weak and there is a nonresonant driving force. it dies out as time increases. There are obviously many variants depending on the strength of the damping and the frequency of the driving force. This part of the solution is called the transient part of the solution. Consider the initial value problem % E % C ð { 7 % E % C { 7 ) g ó Î ò { d ð { " d ð ¤ ¥ ¥ Ó £ Ó ò d U ò õ a a g ô õ ' ' £ ô ~ g ' £ { { g ð { We ﬁnd that 7 ) ' ` ó Î ò ¥ d ) ' ` ò õ ô ¥ µ j ¦ Thus % ¤ d £ ' f Applying the initial conditions gives U " ¥ 7 g ¥ and so the solution is E ) g ó Î ò g " d ) g ò õ ô g C " f ) ' ` ó Î ò a d ) ' ` µ j ¦ ¥ Ó g { This is a combination of oscillations. The term E ) g ó Î ò g " d ) g ò õ ô g C " f µ j ¥ Ó E ) g ó Î ò g " d ) g ò õ ô ! g a C ' ' a " ` ' f a ' ' ` ) ! ' ` ! a ' ó ~ Î ! ' ò g a " ` ' ¤ f d a ' ) ' ¤ ` ! ! g f { µ j d ¦ from which we ﬁnd { { and . Hence ) g ó Î ò ) g ó Î ò 7 Ê " ¤ f ' d g £ g ¤ ' É d f ) g £ g ò õ ô Ê g f É ¥ ¦ £ Ó a ' g { Substituting into the differential equation gives ! ) g ! ó Î ò ¤ d ) g ò õ ô .6.3 OSCILLATION PROBLEMS The case where friction is present This is the ﬁnal case we consider in this chapter.160 8. It is negligible after about 20 seconds. Since the term ) ' ` ó Î ò a d ) ' ` ò õ ô µ j ¦ is damped by the factor.
13. This is called the steady state part of the solution. The next example illustrates a case where the damping is much stronger than in the case above. Hence ) ó Î ò ' " ) ó Î ò E ' " ¤ ' d ¤ £ ' g d f C £ g d f ) ò õ ô E ¤ g £ d £ g g g C g { Substituting into the differential equation gives % E % C ð { 7 % E % C ÿ { 7 ) g ó Î ò { ) ó d Î ð ò { ¤ ° d d ) ð ð ò þ { Figure 8. and so the solution is ) ó Î ò ' d ) ò õ ô g f ÿ ) ' ó Î ò ¥ ' d ) ' ¥ ¤ ò õ ô g ¥ f ¥ þ ¦ We solve these equations to get ¥ Ó { d { { and . We have 7 ) ' ó Î ò ¥ d ) ' ò õ ô ¥ ¦ ¥ Ó { % E % C ð { 7 % E % C { 7 ) ó Î ò ' " { U d ð { g d ð ð { Thus % ¤ g d £ ' ) ó Î ò ' d ) ò õ ô g f Ê ) ' ' ó Î ò ~ " ' " f f ) ' g ò õ Applying the initial conditions gives ô g É ¦ ¥ Ó { d { { .13: The solution to Á % " õ ô £ r g { ¨ " " % f .FORCED OSCILLATIONS 161 is an oscillation at the driving frequency and dominates once the transient has died down. The solution is plotted in Figure 8. Let be a given initial value problem.
but now the damping factor is stronger than in the previous case and the transient part dies out much more rapidly. The solution is shown in Figure 8. r g d % d % r . where r is a constant. We have to make an intelligent guess about the form of the particular integral . EXAMPLE 8.14: The solution to ¦ Á g " " g g ' ' % ¨ f f f g .6.14. The contribution of the transient part is only apparent during the ﬁrst second and is very small. ¥ Ó This structure of transient and steady state solutions applies in any situation apart from very small damping when an irregular transient solution or resonance may persist.162 OSCILLATION PROBLEMS This is again a combination of a transient part and a steady state part.33 Solve the equation Here we take { ) ó Î ò ¤ d ) ò õ ô ¤ g { E ) C RHS 7 U g r g { so that . 8.4 More on undetermined coefﬁcients The method of undetermined coefﬁcients (page 155) can be applied to the equation E ) C { ¤ d ð { d ð ð { r in cases where is not of the form . This is illustrated in the worked examples below. We substitute this into the given equation to get LHS % E % C ð { 7 % E % C { 7 ) ó Î ò U ' " { g { U d ð d ð { ' { g d ð d ð ð ð { { Figure 8.
34 Solve the equation The complementary function is the solution of and is easily seen to be .35 Solve the equation ¤ ¥ RHS 7 Ó r EXERCISES 8. we take the given differential equation to get r ' d % ¤ ¥ £ U ð { ' d ð ð { LHS RHS 7 U EXAMPLE 8. g Ó d ¤ d g { § ¥ d Ó £ { { g ~ Ó Ó r g { g ~ " r so and . Otherwise. The general solution is d { § Ó { 2 r % g { { g g d ð d { % ' ¥ d ð Ó ð { ð ¤ ' { The complementary function is the solution of . we will run into trouble. 2. because this is and substitute this into just a special case of the complementary function. If we attempt to take as we did in the above example. 2 g ó Î ò { d ð ð { 2 ò õ ô r { g f ð { d ð ð { 1. ﬁnd the general solution.FORCED OSCILLATIONS 163 EXAMPLE 8. Instead. ﬁnd a particular solution satisfying the initial conditions. so that the general solution is { d { { ¥ Ó ¤ d ¥ and is easily seen to be d d g ð ð { ¥ Ó £ r g { Ó £ . The general solution is Ó r ` d Ó r U LHS d § Ó { The complementary function is the solution of and is easily seen to be We take and substitute this into the given differential equation to get % ð { ' d ð ð { Ó r g { £ 2 ' U d Ó r ¤ d ð g { { § ' ¥ d Ó d ð £ { ð { { ' ~ 2 U 2 r g { ' ~ U r so and .6 Solve the following differential equations. Where initial conditions are given. .
18. then we can differentiate the above equation to get ) ¨ ~ ¨ E ) C 2 If denotes the charge on the capacitor at time . ( . ﬁnd its position at any subsequent time . for what values of does resonance occur? E ) C 2 U d ð 2 9 g d ð ð 2 E ) C 2 16. If a force of is now applied to the mass. This is known as an 6 ¥ ® E ) C ¥ ® 6 ) % r a ± ) g ó Î ò % g E ) C ð 2 " ¥ d 2 ) ¨ ¨ 6 d 2 ) ¨ ¨ ® ® 6 If .164 OSCILLATION PROBLEMS 17. 6. ¦ E k C 7 % E % C ð { 7 % E % C { 7 ) ò õ ô { d ð ð { 9. " E % C ð { 7 g E % C { 7 2 ó £ Î ò U { d ð ð { 7. What does this mean physically? r ) r ó Î ò r E ) C % 9 (a) If and . 11. A mass of 30 kg is supported from a spring for which N/m and is at rest. It can be shown . which . ¦ ! ð 2 U d ð ð 2 § Ó ` 2 g d ð 2 ' f ð ð 2 12. Give some examples of common phenomena in which resonance is important. contains a time varying voltage source circuit. ) ò õ ô 2 d ð ð 2 2 g ó Î ò ' { r d ð { g d ð ð { 5. 4. Electrical circuits can be modelled by linear differential equations. and are constant. a capacitance and a resistance . % E % C ð { 7 % E % C { 7 ) ó Î ò £ { d ð ð { 8. The position of a mass suspended from a spring satisﬁes the differential equation ¥ Ó r 2 g d ð 2 U d ð ð 2 g { ` f ð ¥ 2 f ð ð 2 14. an inductance . 13. then the current that the charge obeys the differential equation E ) C ( " ¥ ) d ( ¨ 6 d ( ¨ ® ( E ) C ( ¥ ® 6 ) E ¨ ) C 2 Ñ Ï Î Ð Í ) ¦ ¨ E ) C ! E ) C r 9 g (b) If and . 2 ' ó Î ò ' { d ð { f ð ð { 2 ó Î ò ð { f ð ð { 3. Ó ` 2 d ð ð 2 ¦ a " 2 ' d ð 2 U f ð ð 2 10. 15. ﬁnd . Consider the circuit below.
[Hint: In the equation volts. Assume that in the above and current at any time if: 6 ) circuit. Find the charge volts. ¥ . where is a constant] ¥ ® 19. henry.FORCED OSCILLATIONS 165 6 % E ) C ´ ¥ % " § ! 20. An circuit has henry. for which the circuit is critically damped. for . farads and volts. the initial values of and are both zero. take henry. g % ® ® 6 farad and ) ò õ ô E ) C r ¥ " ® U ¥ 6 (c) ohms. . farads and . henry. Find the resistance ) g ó Î ò E ) C g " ~ " ¥ g ® % " 6 (b) ohms. farads and ` g E ) C ' " ~ " ¥ ± " ± ® g U 6 (a) ohms.
If we want to analyse the motion around the curved in Figure 1. The car has an acceleration as it traverses ~ ò õ ô 6 ~ ó Î ò ¤ ~ ó Î ~ ò º 6 ¯ ~ ò õ ¥ ô ¥ ¤ . 9. In this chapter. we only dealt with the motion on the vertical part of the tower. There are two forces on the car—gravity and the normal reaction . that is. the mathematics we have developed is inadequate part of the track ( to do this.1 ANOTHER PROBLEM ON THE TOWER OF TERROR Let us concentrate on the curved part of the motion in the Tower of Terror (Figure 9. These were formulated in terms of differential equations.1). which we were able to solve after developing suitable mathematical methods and concepts. we did not consider all aspects of these problems. This angle will be a function of and the problem is to ﬁnd this function. we will begin to develop methods which can be used to solve this problem. We can specify the position of the car on this arc by using the angle shown in the diagram. These new methods will also provide an alternative way of solving the previous problems. We can resolve the force of acting tangentially to the circle and a component acting gravity into a component perpendicularly to the circle.CHAPTER 9 INTEGRATION In Chapter 1 we began with three problems. g ~ ~ % ~ ) ) ~ Figure 9. to ﬁnd the rule for calculating values of from values of . However. For instance. where . in the Tower of Terror. where is the mass of the car.2 of Chapter 1). We assume that the curved part of the track is an arc of a circle of radius metres.1: At the bottom of the Tower of Terror.
9. the frame must apply a reaction force . it is easier to go back to the original three problems presented in Chapter 1 and look for alternative ways of solving them. We solved it by a power series method to get the solution .168 INTEGRATION the circle. so for values of close to zero. This comes about because it is being forced to move in a circle rather than a straight line and because its rate of traversing the circle may be changing. so that the net radial force on the car is . Using Newton’s second law of motion we have ~ ð ó ~ ð Î ð ò ~ ð ~ 6 ~ ó Î ò ~ ò ¤ f õ ô º ) ð ¨ ¨ ~ ð ~ d ~ ò õ ô º or where . If we were to try to solve this equation by assuming a power series for of the form ) ¦ ¦ ~ ¦ d § ) r d ) r d ) r d r ~ ~ ~ 7 % ~ ó Î ò d ð ð ~ for . Rather than leaping into the unknown by considering this problem now. This is the differential equation we have to solve to ﬁnd as a function of . This net force keeps the car is thus on the circle and prevents it ﬂying off on a tangent. the lowest portion of the curve. We can then return to the present problem and use our new methods to solve it. we would run into difﬁculty with the term ç ¦ ¦ ¦ d § r ) r d ) r d ) r d r æ ó Î ò ~ ó Î ò The basic problem is that the differential equation is not linear and so our methods for linear equations are not applicable. that is . . Consequently. then ~ ~ ¨ ~ § § ó Î ò f ) ¨ ~ and so ~ (9. is not zero and we The initial condition we used was can write the differential equation as { 2 " { 2 ¨ ¨ " " { E % C { Ó { º If is the radius of the circular arc and is measured in radians. Since inﬁnite series cannot be easily used to solve this problem. we shall develop an alternative approach. The total reaction force supplied by the frame to the radial component of its weight. We are now going to get this same result by a different method.2 MORE ON AIR PRESSURE The initial value problem " E % C { 7 { { 2 ¨ ¨ arose in our discussion of the air pressure problem in Chapter 6. Since the radius of the as well as a reaction circular arc does not change. This will take some time and we will not be able to complete the solution to the problem of ﬁnding the motion along the circular arc until Chapter 13. but does not affect its speed and plays no role in our problem. Let denote the distance along the curve from . although it will take us most of this chapter and the next to do this. It can be shown that this acceleration can be resolved into a radial component and a tangential component . we need only consider the tangential force .1) .
. how do we ﬁnd a function such that ? The answer to this question will lead us into primitive functions (or antiderivatives ) and integrals. how do we ﬁnd ? To answer this question we will need to understand inverse functions.3) 7 that is. The solution and it is difﬁcult to see how to disentangle it. If we are given a function . The composition 3 satisﬁes .2) We now have a new problem: ﬁnd a function whose derivative is 1. As we shall later show. Such functions are easy to ﬁnd. Suppose that we are able to ﬁnd a function whose derivative is given by . However. ¤ d 2 E 2 C E { û C At this stage we know the following: 2 ~ " E 2 C ð û is an unknown function whose derivative satisﬁes ¤ d 2 E 2 C E { û C { û û .MORE ON AIR PRESSURE 169 The key to the new method is to use the chain rule on the lefthand side of this equation. If we know the function and the composition . { { û û 3 ð û û 3 { û û ¤ d 2 for any constant . These will be discussed in Chapter 10. We have ¨ and. unknown function in the composition we have identiﬁed two problems that need to be explored in greater detail. " E ¤ d 2 C ¨ (9. any function with derivative 1 must have the form so let’s assume this for the moment.2) as ¤ d 2 ç E 2 C { æ û ¤ 7 " " " E E E g 2 C ' d 2 d ¨ 2 C 2 C 2 2 ¨ ¨ 2 ¨ 2 ¨ ¨ ¨ ¨ 3 . E 2 C { that we want is tied up with the It is not obvious that we have achieved anything. which we discuss in the next section.1) can be written ) ~ " E ) C ð û 7 " { 2 ¨ ¨ E { C ð û û or " ç E 2 C { C û æ (9. Then we can write equation (9. Then equation (9. in general.
and so % E 2 C f E 2 û C E 2 C ð f E 2 C ð û ç E 2 C f E 2 C û æ ¨ ð ð û Thus . where is a constant. Knowing results for derivatives enables us to ﬁnd primitives: ¤ E 2 C û 2 E 2 C " f k 7 ¤ ¤ d Ú ¤ 2 2 ¨ ¤ 2 ° 3 EXERCISES 9. Suppose and are primitives of . primitives and the process is intrinsically harder than differentiation. then û for some constant . an antiderivative or an indeﬁnite integral of . A given function has many and are all primitives of . is a is a primitive of and so on. In other words. ¤ ¤ d THEOREM 9. Note that Mathematica does not put in the constant .1 If and are primitives of the same function. then so is for any constant . 2. For example.3 Find primitives for the functions in Exercises 1–6. it is easy to see that we can add a constant to any primitive function to get a new primitive function.3 INTEGRALS AND PRIMITIVE FUNCTIONS Let be a given function. If is a primitive of . A function such that is called a primitive.170 INTEGRATION 9. we have found all of them. ¤ d û 2 g U U d d 2 2 ó Î ò û f ¤ " ¤ d d d 2 r 2 2 ò r ¤ õ ô ó 2 Î d ò " r " " 2 r ß f ¨ Ó d " r 2 2 ò 2 ¨ õ ¨ ô 2 2 2 ¨ r r ò ó ß õ Î ô ò Ó ° ° ° 3 3 3 ¤ . 2 g ó Î ò d ° Ó E 2 C " d 2 g d § 2 E 2 C 1. Finding is the reverse of the problem of differentiation. Then . û 2 ð û û û This theorem is a consequence of the result (which we stated on page 100) that a function with zero derivative on an interval is constant on the interval. a constant is zero. Thus . There are no comprehensive rules for ﬁnding primitive of 1. and hence . There is a special notation used for primitives. The others differ by constants. In fact. The ﬁrst point to notice about primitives is that they are not unique. if is a primitive function of . Since the derivative of primitive functions. The notation has arisen for historical reasons and we shall have more to say about it later. A consequence of this theorem is that once we have found one primitive of a function. we have the following result. we write û ¤ d E 2 C E ¤ 2 C d û E 2 C û 2 ¨ E 2 C ° ¤ ¤ E 2 C f E 2 C û This is read as the (indeﬁnite) integral of with respect to is plus .
AREAS UNDER CURVES 171 9. It is possible that there is some complicated rational function which when differentiated by chain rules. Determine E g f 2 C E 2 C E " f 2 C 2 { 5. In fact it can be shown (with some difﬁculty) that there is no function deﬁned by an algebraic formula . 3 In this chapter we will show you how to use integration to produce functions having a given derivative. be a of . One of the most fundamental of these arises as follows. Let ﬁxed point in the domain of a positive continuous function (Figure 9. there is a way of deﬁning a function such that .4 AREAS UNDER CURVES In Section 9. A function satisﬁes and . 6. This is unlikely.2: The area under a curve ) E ) C £ Calculus is full of surprises. Can we in fact determine a convenient expression for ? Certainly the primitive of is not a power function since 2 ~ " E 2 C ð û û 7 " f k û 7 ¤ û d " Ú d ¤ 2 2 ¨ ¤ 2 ° where is an arbitrary constant. . So is there any such function at all? Around the period 1650–1670. This is done by a process known as deﬁnite integration and provides a third way in which to deﬁne functions. Usually the process of differentiation results in a more complicated function. Recall that we have so far deﬁned functions in one of two ways: The simplest types of function were those deﬁned by algebraic formulas such as " d 2 ' d 2 3 û û 2 ~ " ¤ 2 ð ~ " û Functions deﬁned by power series arose when we solved differential equations. ¨ Figure 9. If we are given any continuous function . § " 2 E 2 C 2 ' ó Î ò d 2 ô w ò E 2 C 3. Consider the area û P E 2 C g { ) ' d § 2 E % C ´ û ³ ² E 2 C ± " d ) ' ) E ) C ð û û 7. end up as .2). but in general any power series deﬁnes a function whose domain is the interval of convergence of the power series.1 we assumed it was possible to ﬁnd a function whose derivative satisﬁes . Sir whose derivative is Isaac Newton and Gottfried Leibniz developed the ideas which led to one of the great discoveries in calculus. after much cancelling. product rules and the like may. 4.
3). (Figure 9. The surprising fact is that . so that its area is point . ð E £ ) f 2 C f { ) 2 £ f 2 2 f ) ) E 2 C 2 E ) C £ and that .2 as in the previous example. where is any point in the domain of which satisﬁes . but differs from the primitive in the previous 2 E 2 C Figure 9. ) 2 { ) ¨ 2 £ ð 2 £ 2 ¨ ) ) E 2 2 C E ) C £ ð ) £ 2 E ) ) C £ 2 ) ) w ) 2 E ) ) C £ ) E 2 C { £ ) ) EXAMPLE 9.4). the line and the vertical line through a Let point on the positive axis is a triangle with base and height (Figure 9. the axis. EXAMPLE 9. the axis and the vertical lines through and . The region bounded by the axis. In this case.3 .172 INTEGRATION the region bounded by the graph of . so that its area is .3). the line and the vertical line through a Let on the positive axis is a triangle with base and height (Figure 9.3: Regions under 2 P E 2 C ) ´ ³ ² ± ) 2 P E 2 C ) ´ ³ £ ² ± £ . and this does not appear to have anything to do with the area under the curve. Hence example by a constant. It is certainly not a result we might have expected. The connection between area and primitives is not just mathematical entertainment. and the two vertical lines and . It is easy to see that ) 2 ) 2 7 ) f ) E 2 ) C £ 2 2 E 2 { C ) E ) C £ EXAMPLE 9. which can be physically interpreted as the slope of a curve at a given point. ð is again a primitive of .1 . Why should this be so? A function and its primitive are connected by differentiation. It also turns out to be useful. so that is a primitive of . Here we have . The region bounded by the axis. Before proceeding with the full details we shall consider a few simple cases to illustrate the connection between primitives and area. We loosely refer to as the area under the curve between and . Notice that in this way we can deﬁne a function (denoted by ) if we let the ouput number be for the input number . that is . it is which is a primitive of . but this time we consider the area Let us take the same function of the region bounded by . There is an even greater surprise in store.
In it. the parabola Let and the vertical line through a point on the positive axis (Figure 9. We can solve this problem by making use of a result which has been known for over 2000 years and which was ﬁrst proved by Archimedes in about 250 BC1 .4: A triangular region under 2 f E P 2 C § 7 ) § ) r ) § § ´ ³ ² ± £ ¨ µ £ . so is a primitive of .5: The area under a parabola 1 This discovery is documented in a remarkable piece of work entitled Quadrature of the parabola. The signiﬁcance of his ideas were not realised until the seventeenth century when the development of the calculus occurred. Using this result. which can be found in the Encyclopaedia Britannica series. We see that . 2 f E 2 C Figure 9. we see that 2 2 ) r f ) r ) E ) C E 2 ) £ C £ ) r § ° 2 E 2 C 2 ( » © { Figure 9. We wish to ﬁnd the area of the region bounded by the axis. Archimedes performs what is essentially a numerical integration to reach his goal of ﬁnding the area of a parabolic segment.AREAS UNDER CURVES 173 EXAMPLE 9.4 .5). i 2 E 2 C ( P ) ´ ³ ² ± ) ( r ¨ » ( © ð ) r since . Interested readers may like to read this work. Archimedes found that the area of the parabolic segment in Figure 9.5 is . Great Books of the Western World (volume 11).
174
INTEGRATION
The key observation to be made in all of these examples is that the area of the region between the graph of the given function and the axis is, to within a sign, also a primitive for the function. This is not a coincidence; rather it is one of the most fundamental facts about the calculus and one which will need to be carefully investigated.
9.5 AREA FUNCTIONS
Let be a given function and let be a ﬁxed, but arbitrary, point on the axis in the domain of . We call a reference point. We want to deﬁne another function by making use of the area of the region bounded by the graph of , the axis and the vertical lines through and any other point in the domain of , with . We shall have to be careful about signs when doing this. We begin with for all in the domain of . Then (Figure 9.6) the most straightforward case by assuming we deﬁne by where
C 2 û ) )
The area (and hence the function value ) is a number which can be calculated for each choice of . The calculation is not always easy and we shall have to investigate methods for doing it, but for the moment let’s accept that it can be done.
) E ) C û ) E ) C )
Now let us take the case when for all in the domain of (Figure 9.7). As before, choose a reference point and let be any point in the domain of satisfying . Let be the
E ) C £ ) w ) 2 % u E 2 C ) )
E
)
C
f
E
)
C
û
û
Figure 9.7: Deﬁning a function
by
E
)
C
£
P
E
E
)
2
C
C
û
{
û
Figure 9.6: Deﬁning a function
£
by
£
)
P
E
2
C
)
is the area of the region under the curve between
¨ E
and .
)
)
2
{
)
7
E
)
C
%
£
w
´
´
³
³
E
E
²
²
)
2
C
C
±
±
û
2
2
)
)
¨
w
µ
)
E
)
C
û
)
£
AREA FUNCTIONS
175
Finally, we take the case where changes sign one or more times for . In this case, some parts of the region bounded by the lines , , the graph of and the axis will lie above the axis, while other parts will lie below the axis. We attach a positive sign to the area of each part of the region which lies above the axis and a negative sign to the area of each part of the region which lies below the axis. The function is deﬁned by letting its output number at the point be the sum of these signed areas (Figure 9.8).
E ) C û 2 ) u 2 u ) ) 2 2 ) û 2 2 E 2 C 2 2 )
In summary we can use the area of the region bounded by the graph of a function , the axis and two vertical lines to deﬁne a function . In Examples 9.1–9.4 this function satisﬁes the condition (Exercises 9.5). There are two matters which we need to address: How do we calculate the area of the region under a curve in the general case? We need to be able to do this in order to calculate function values of .
ð û û 2 û û û 3 ð û
Does the function
3
always satisfy
?
The answers to these questions emerged with the discovery of the calculus by Newton and Leibniz in the middle of the 17th century. The answer to the ﬁrst question is conceptually straightforward, although the calculations may sometimes become unwieldy. We shall address this question in the next two sections. The answer to the second question is given by the fundamental theorem of the calculus, which we shall discuss in Section 9.8. EXERCISES 9.5 1. Construct the function described in the text of this section for each of the examples 9.1–9.4. Verify that in each case. 2. Use Archimedes’ result for the area of a parabolic segment to show that the area of the region , the axis and the vertical lines and is given bounded by the graph of . Is a primitive of ? by
) 2 g 2 2 2 ' E 2 C £ ! f § ) E ) C £ û ð û
d
f
E
)
C
û
û
Figure 9.8: Deﬁning a function
by
)
)
P
£
E
2
C
£
{
)
¹
E
)
¶
C
£
f
E
)
C
û
·
¶
)
2
area of the region bounded by the we deﬁne by
E ) C û
axis, the graph of
and the vertical lines through
and . Then
¨
176
INTEGRATION
9.6 INTEGRATION
In the previous sections, we have shown how it is possible to construct a function by using the area (with an appropriate sign) of the region bounded by the graph of a continuous function , the axis and , where and are in the domain of . The ﬁrst problem to be dealt and the lines with is to determine a method of calculating the area of a ﬁgure with one or more curved boundaries. The basic idea is to approximate the area by simple shapes whose area can be calculated and then take a limit. This idea is a very ancient one. It was used by Archimedes to calculate the area of the parabolic segment we considered earlier. 9.6.1
2 û
Positive functions
7 r g 2 % w E 2 C 7 r
and suppose that for all . We wish Let be a function deﬁned on the interval , the axis and the lines and to determine the area of the region bounded by (Figure 9.9). The simplest way to approximate the area of the region is by rectangles. Divide the interval into subintervals where (
7 " 9
Figure 9.9: Approximating the integral by a Riemann sum the areas of the individual rectangles shown in Figure 9.9. Such sums approximate the area under the curve and it seems intuitively clear that the approximation converges to the exact value of the area as we take more intervals in such a way that their number increases without limit and such that they all get arbitrarily small. This is true if is continuous. However, it is quite difﬁcult to prove this, since
E
2
P
C
{
º
³
2
°
2
º
°
§
2
2
º
§
2
2
º
2
2
7
º
2
7
g
7
r
"
¨
¥
2
f
2
2
where
,
. Such a sum is called a Riemann sum. It is the sum of
2
7
E
¥
¤
2
2
and . The intervals need not be of the same length. In each subinterval ), select an arbitrary point and form the sum
f ¥ ¤ 2 C E º ¤ 2 C d ¦ ¦ ¦ d E 2 f 2 C E º 2 C d E 2 º f 2 2 C E º 2 C d E 2 f 2 ¤ C 2 E º 2 C
7
r
2
r
2
¤
2
¶
7
E
º
¤
¤
2
2
2
C
7
§
¥
d
¤
E
¦
2
¦
§
2
7
C
¦
d
7
§
2
{
2
¶
)
7
E
º
§
2
2
C
7
)
d
2
7
2
2
¶
E
º
)
2
C
7
2
2
d
¶
E
2
º
2
¶
C
E
)
º
2
C
¤
÷
2
9
7
r
E
7
g
¶
C
2
INTEGRATION
177
we have to consider all the possible ways in which the intervals become small and increase in number. We shall omit the detail on how to do it in a mathematically correct way and simply write
X ¤ 9 W °
(9.4)
. The The expression on the lefthand side is called the deﬁnite integral 2 of on the interval numbers and are called the limits of integration. We call the lower limit and the upper limit. The process of using Riemann sums to compute integrals can be simpliﬁed considerably if we into subintervals take intervals of equal width. Begin by dividing the interval
7 r 7 ¤ 2 7 7 r r ¥ ¤ 2 7 7 2 7 2 7 2 7 2 r
of width
7 r f 2 ¶
where and . There are now two obvious ways to approximate the area by rectangles. Firstly, we can use rectangles of height and base ,( ). This gives the following approximation to the area under the curve
2 ¶ E ¥ ¤ 2 C d ¦ ¦ ¦ 7 d 2 ¶ 7 E g 7 2 " C d 2 2 ¶ E ¶ 2 C E E C ½ ¥ 9 2 C ¤ 2 r 2
The particular Riemann sum is called a lefthand sum , because we have used the value of the function at the leftend of each interval. This approximation is illustrated in Figure 9.10.
E C ½ 9 ¨
Figure 9.10: The integral approximated by a lefthand sum
¾
2 The
integral sign
was ﬁrst used by Leibniz on October 29, 1675. It was an elongated “S”, for the word “sum”.
E
2
C
Alternatively, we can calculate a righthand sum by using the rectangles of height
E
2
P
C
{
ÿ
E
E
¥
¤
¥
2
C
2
f
d
¦
³
2
¦
C
2
¦
E
º d 2 ° C E 2 2 C ÷ § 2 2 d E ¶ Ñ Ð Ï ¼ 2 2 Î Ð C Í ¤ » þ 2 r 2 f ¨ r E 2 C ß
(9.5)
and base
178
¶
INTEGRATION
Figure 9.11: The integral approximated by a righthand sum will converge to a number which is equal to the area of the region in question and equal to the limit of the Riemann sum (9.4). Thus for any continuous positive function , the integral
°
may be calculated by taking a limit of lefthand sums:
r f °
(9.7)
ÿ E
or by taking a limit of righthand sums:
r f °
(9.8)
9.6.2
General functions
7 r
and construct the limit of Riemann sums
X ¤ 9 W
7
7
E
¤
2
¥
7
2
¥
f
¤
2
2
7
C
E
º
2
7
C
7
2
r
7
If
is not necessarily positive on the interval
2 7 2 7 2
, we can still divide
into
subintervals
¡
E
C
¿
9
¡
E
E
C
2
P
C
½
9
ÿ
E
¥
¤
¤
{
2
C
2
C
d
d
¦
¦
¦
¦
¦
¦
d
d
E
E
³
2
2
C
2
C
°
d
2
2
d
¨
E
E
E
2
2
§
C
C
2
C
2
2
þ
þ
¶
ß
÷
2
It seems intuitively clear that as we allow
¨
to increase, both of the sequences
2
ÿ
¶
E
E
¤
¤
2
2
C
C
d
d
¦
¦
¦
¦
¦
¦
d
d
E
2
2
¶
C
E
2
C
d
E
d
2
C
2
þ
¶
E
r
Ñ
Ð
Ñ
2
Ï
¼
Ñ
Î
Ï
2
Ð
f
Í
Î
C
Ï
Ð
Í
¤
Î
Ð
»
¤
Í
¤
r
Ï
Î
Í
E
2
2
¨
C
¨
E
E
¿
2
C
2
9
C
ß
ß
E
7
7
g
7
"
,(
2
, as illustrated in Figure 9.11. This gives the approximation
(9.6) and
INTEGRATION
179 ,
As before, we call this the deﬁnite integral of on the interval . However, the interpretation of the integral is now different. In this case, some parts of the region bounded by the lines , , the axis and the graph of may lie above the axis, while other parts may lie below it. We attach a positive sign to the area of each part of the region which lies above the axis and a negative sign to the area of each part of the region which lies below the axis. The integral is then the sum of these areas, taking the sign into account. This statement follows directly from the way in which the for all , then the limit of the sum Riemann sum was constructed. If, for example, in equation (9.4) will be negative. in There are some minor points to clear up. In the above discussion, we have assumed that . If this is not the case, we simply deﬁne for by the integral
r w 2 r 2 r u 2 2 ¨ 7 E r 2 C g 7 ß 2 r À 2 % ß u ° E 2 C 2 ° 2 ¨ E 2 C ß À 2
Also, for equal upper and lower limits, we deﬁne
% 2 ¨ E 2 C ß °
9.6.3
Some properties of integrals
There are some other properties of integrals which we now state. First, there are the results for sums and multiples:
° ° æ °
and
° °
where is any constant. These results can be proved in each case by expressing both sides either as the limit of a lefthand sum or as the limit of a righthand sum and showing that these limits are equal. There are no results for integrals of products and quotients which apply in all cases. , then If we take a point with
° ° u ¤ u ° r ¤ ¤
If is positive, then this result amounts to the addition of areas, but the same result holds whatever the order of , and and whatever the sign of . The use of lefthand or righthand sums is inadequate to prove this result. A correct proof would need to use more general Riemann sums.
¤ r
7
where is an arbitrary point in the th subinterval. As long as the function is continuous on it can be shown that such a sum will converge and once again we denote the value of this sum by
r ° ð 2 º ) 2
2
¨
E
2
C
2
¨
E
ß
2
C
2
7
d
¨
2
E
¨
2
2
E
¨
C
E
2
d
C
2 C 2
2 ¨
¨ ß E
E ß 2 f
2 C ¤
C
ß 2 ß 2 2 ¨ ¨ ¨ ß E E ç 2 E 2 C C 2 2 C ¨ ¤ E ß 2 d ß C E 2 C ß ß
180
INTEGRATION
We can now return to the area functions we were discussing in Section 9.5. For any continuous function with domain we can deﬁne a function by
û ¦ ° û 7 r
The discussion in Section 9.5 shows that the number can be interpreted physically in terms of the area of a region bounded by the the graph of and the vertical lines and . Functions deﬁned in this way will play a major role in the remainder of this book and it is important that you understand this process. In general, integrals cannot be calculated exactly except in special cases. This is because no formula can be found for the Riemann sum and so there is no way of calculating the limit. We can compute approximate values to any required degree of accuracy by choosing sufﬁciently large values of . These calculations are usually very tedious to do by hand and a computer is needed. EXAMPLE 9.5 Evaluate
2 ¨ " ° ) 2 r 2 E ) C û
If we choose ( ) we obtain 0.706 for the approximation while if we choose ( ) we obtain 0.696. The sequence of values appears to be converging, but very slowly. There are various ways to improve the approximation and we shall deal with these in the next section. In some special cases, it is possible to use formulas such as
¤ W % i % " r % % ( % g " % % (
(9.9)
E
(9.10)
E
(9.11)
to evaluate integrals explicitly by ﬁnding an expression for a lefthand or righthand sum and then taking a limit. The above formulas can be proved by the use of mathematical induction. EXAMPLE 9.6 Evaluate
7 2 ¨ E 2 C °
2
E
2
C
where
.
"
%
¥
(
"
"
a
d
%
g
E
C
E
"
Ê
"
"
d
"
d
"
d
C
C
d
C
¦
¦
°
¦
d
§
g
±
±
±
"
"
¤
¤
÷
÷
÷
d
W
W
"
"
"
§
d
%
d
d
d
"
¦
%
¦
¦
¦
"
¦
¦
¦
"
¦
¦
É
d
d
d
"
'
§
%
'
'
d
g
d
d
2
§
d
¨
g
g
"
"
2
d
d
§
"
"
g
°
approximately. Here , Then
" r
. We choose
, so that the width of the subintervals is
)
r
7
2
¨ 2
E
2
C
ß
E
)
C
.
INTEGRATION
181
Thus
" °
Note that this gives Archimedes’s formula quoted earlier. EXERCISES 9.6 Evaluate the sums in Exercises 1–9. 1.
C ÷
10. Determine the limits of the following sequences.
d ¦ ¦ ¦ d ' d g d "
(a)
"
U
7
%
r
7
"
d
2
'
E
2
C
11.
r
2
In each of the following exercises, ﬁnd the area of the region under the graph of by expressing the integral as the limit of a sum.
2
§
d
¦
¦
¦
d
°
§
'
d
§
g
d
§
"
(c)
Ô
Õ
d
¦
¦
¦
d
§
'
d
g
d
(b)
Ô
E
"
d
±
C
÷
E
f
±
C
÷
Õ
Õ
§
Ô
±
7.
³ ÷
8.
9.
from
E
±
'
d
±
C
°
÷
g
W
E
2
g
f
§
2
C
¤
÷
)
W
E
g
f
1
C
E
"
d
1
C
4.
÷ g
5.
6.
E
"
f
±
C
÷
÷
W
E
'
d
C
2.
÷ ¤ E
3.
Â
Ê
ÿ
Â
þ
W
Ê
Â
~ ç d
" ¦
¦
d
¦
¦
Â
d
¦
E
¦
d
¦
"
¦
¦
Ê
d
d
g
d
É
'
g
C
'
`
E
d
"
~
d
d
d
g
g
Ê
"
C
d
d
¤
É
2
"
"
§ æ
É
É
§
§
"
"
"
"
W
Á
Á
Á
Á
W
~
" g Ñ Ñ Ñ Ñ
7 Ï Ï Ï Ï Î Î Î Î
% Ð Ð Ð Ð
Í Í Í Í " ' ' ¤ ¤ ¤ ¤
Solution. We divide the interval with , , ,
2 ~ " ° 2 " % r
into ,
subintervals of width , we ﬁnd
. Then using equation (9.8)
2
2 ¨
¨
2 2 " f ± ' µ W W W
to
as shown in Figure 9. r 7 % r 2 ' E 2 C 12. we see that the area of a typical trapezium in the area is found to be Figure 9. Area (9. but we shall not prove it here.7 EVALUATION OF INTEGRALS In this section we shall assume for simplicity that all functions are positive. Trapeziums are foursided we divide the interval ﬁgures with a pair of parallel sides. 9. 2 ¨ E 2 C ß À ~ " E r f C ç E E ( 2 C d r d C E ( ¥ 2 C æ ( ¨ 2 r 7 r 2 Figure 9.1 The trapezoidal rule It would appear from a consideration of Figure 9. To get an approximation to into subintervals of width and approximate the region under the curve between and by trapeziums. r 7 " r 2 ' E 2 C 13.12) . The formulas we get will hold even when this restriction is lifted.12.12 is . where . We ﬁnd that æ ( æ ( æ ( 2 is an approximation to Equation (9.7. Using this result.12. U 7 " r 2 f % " E 2 C 15.12) is known as the trapezoidal rule. Then .12 that a better approximation to the area under a curve would be obtained if we used trapeziums instead of rectangles. .12: Approximating a region by trapeziums ¤ t ¤ t Let E 2 C ß À denote the sum of the areas of all such trapeziums.182 INTEGRATION 9. In the notation of the righthand diagram in Figure 9. E d ç E r ¤ C 2 ( C d E ÿ E ¤ ¥ 2 ¤ C 2 C d E g ( ¥ d ¤ ¦ 2 ¦ r C ¦ g d ç d E P ¦ ¦ 2 ¦ C d ° E d 2 E 2 C 2 C g § 2 d g E ` d 2 C ç 2 E g 2 7 C d E % E 2 C 2 d 2 C E r þ 2 C { g ( 2 E " g f 2 C ¤ t E ¨ 2 C 16. r 7 % r " d 2 g d 2 ' E 2 C 14. The area of a trapezium is easily derived by dividing it into two triangles and a rectangle.
.13). For .7 Approximate 2 ¨ " ° The convergence is far more rapid than for the rectangle approximation.2 Simpson’s rule An even better approximation is to take points three at a time and ﬁt a parabola through them. we ﬁnd ! ' ` % Ê % % " % g " d g " d ¦ ¦ % ¦ g d g g " % " d " g " using the trapezoidal rule for the cases Solution.EVALUATION OF INTEGRALS 183 EXAMPLE 9.7. ¤ 2 7 ¥ ¤ 2 7 7 2 7 2 7 r ¥ Ã 2 2 7 r we divide the interval into an even number of subintervals.13: A parabolic approximation to part of the graph of E 2 P C { Ú Ã 2 Ã 2 be the end points of these subintervals. we pass a parabola through the points and then ﬁnd the area of the region bounded by this and (Figure 9. the axis and the lines 2 2 C 7 2 7 E ¥ E 2 2 C 7 2 C E E 2 ¥ 2 C 7 ¥ 2 C ( ). we have " " % % " . where ¨ is an even integer. and parabola. 9. d % " É g t . Let g ' ` % ß t % % " while for 7 ' ' ` % t % g For . We can use the result E 2 C { Figure 9. 2 and . To approximate 2 ¨ E 2 C ° " Ú f 2 7 2 7 g E 7 E " Ú ¥ 2 C 2 ¡ 7 Ú Ú 2 For a typical triplet .
Denote the sum of these areas by and we ﬁnd. In using approximations such as the trapezoidal rule or Simpson’s rule. that approximation to the integral d E 2 C U d E 2 C g d E 2 C U d 2 E ¨ E 2 C 2 C ß ( À ¤ 9 . § E r f g " C p  E ¤ t C ¡  7 Suppose that for all in r 2 3 is continuous on . for even values of . We shall not discuss these here. The rule which we know by his name was known to Gregory and Cavalieri in the seventeenth century. Then is an ÿ E ¤ 2 C § d E ¥ ¤ 2 C U d E ¥ ¤ 2 C g ( d þ ' g (9. EXAMPLE 9.184 INTEGRATION Equation (9. Let and respecWe can write tively denote the trapezoidal estimate and Simpson’s estimate for ¤ 9 ¤ t 2 7 ¨ r E 2 C ß À ° Thus is the numerical difference between the actual value of the integral and the trapezoidal estimate with subintervals.13) .8 . Take % " We have % r " ' ` % Ê " g d U " d ¦ ¦ ¦ d ' U " d g g " d " U " d " " É " ' % 9 We conclude that the value of the integral is 0. We state the following two results without proof. Then and that there is a constant E ¤ 9 C ¡ d ¤ 9 2 ¨ E 2  C E ¤ 9 ß C ° and. p  E 2 C ð ð  p 7 r ð ð 3 Thomas Simpson (1710–1761) was a weaver who taught himself mathematics. questions regarding the accuracy of these rules naturally arise. E ¤ t C ¡ a U d " ' ¤ t ` % 2 ¨ E 2 C 9 ß ¡  % % "  E ¤ t C and for a U " ' ` % 9 % g ¡  For we ﬁnd 2 ¨ " 2 ° % % " and % in Simpson’s rule to approximate the integral such that ¤ 9 ¤ 9 of Archimedes mentioned earlier to do this. after some effort. By choosing successively smaller values of we look for a sequence of constant output numbers from the application of Simpson’s rule.3 It is sufﬁciently accurate for many practical purposes.6931 to 4 decimal places.13) is known as Simpson’s rule. while is the numerical difference between the actual value of the integral and Simpson’s estimate with subintervals. but there are times when more sophisticated methods are needed. Let be a function with domain . but Simpson popularised it in his book A new treatise on ﬂuxions.
° U " r g p a " % % § E % % " " a ° ® " f 2 § % ! ¨ g % g ' E C " g 2 % C ` %  ß E ¤ 2 t C ¨ ¡  2 7 g 2 7 ¨ " on the interval gives . Consequently. using the above error estimate with ú ú ú ú § g 2 ú ú ú ú g  E 2 C ð ð  2 ~ " E 2 C Solution. . U ` 7 2 ¨ 2 ó Î ò 2. using the result from Example 9.7. if is even and that there is a constant such that 7 2 2 ¨ § ¤ 9 .EVALUATION OF INTEGRALS 185 g EXAMPLE 9. U 7 2 ¨ 2 d " 4. 1.9 Estimate the error in applying the trapezoidal rule with 10 subintervals to the integral " ° Hence. 3. We put . Calculate an approximation to ° using (a) the trapezoidal rule and (b) Simpson’s rule.7 In the problems below. we have " ° EXERCISES 9. " d " 2 ° ¤ t The next two exercises give tabulated values for a function . Use the indicated number of subintervals and round the answers to 4 decimal places. calculate both the trapezoidal approximation and Simpson’s approximation to the given integral. % r g " r 5. and  E 2 C U ° T  ³ E r ° f % ! C " 2 ¨ 2  E ¤ 9 C 7 ¡  r U ° T 7 Suppose that for all in r 2 3 ° is continuous on . so that . Then.
00 2.73 1.0 2.25 % 1.50 1. The creek is 8 metres wide.5 1.80 1.5 differs from its true value by no more that 0. A function is given by the rule U f g a f r a f U ` f r U f U g f a 0 % " % C r 6.0 2. Next. Two students are standing on a bridge over a creek. 0.41 E 1. they measure the depth of water at various points along the width of the creek and derive the proﬁle shown in the diagram.50 1. ¤ t 2 2 ¨ ¨ " 2 2 ° 9. for the ) ¨ ) ó Î ò d " ° E 2 C 8.66 1. 1 2 3 4 5 6 % r g 1. Determine how large integral must be in order to guarantee that the trapezoidal estimate E ' C Use Simpson’s rule to evaluate .2 1. They ﬁnd the average velocity to be 0.5 . ¤ 9 10.25 % " 2 2 C 7 8 0 9 81 10 200 1. First.58 1. they measure the velocity of the water by throwing twigs into the water and timing how long it takes for them to pass under the bridge. .75 2. They decide to measure the volume of water ﬂowing under the bridge in the following way.186 INTEGRATION 1. Use Simpson’s rule to calculate the volume of water per second ﬂowing under the bridge.5 2.0005.5 m/sec. Determine how large must be in order to guarantee that Simpson’s estimate " ° for the integral ¥ % " differs from its true value by no more that . where all depths are given in metres. by using a long rod.87 0 f E 2 2 7.
THE FUNDAMENTAL THEOREM OF THE CALCULUS 187 9. that is E 2 C û f E ¤ ( 2 C û æ " l Ï Î Í Hence Another way of writing this result is ) ¨ E ) C ° ¨ E 2 C E û EXAMPLE 9.2 The fundamental theorem of the calculus If is continuous on the interval and if is deﬁned by ° û 7 r The function is well deﬁned because we can calculate its function values by procedures such as Simpson’s rule. we consider X ° Ú ° W û û æ " Taking the limit carefully as (and this is where we gloss over the detail) we ﬁnd that the approximation becomes an equality. By deﬁnition outline of the proof without going into detail. such that . We will only give an converging to zero.10 Find a function such that " 2 E 2 û û E 2 C E 2 C E ¤ " ( ç E 2 ß f E 2 ¤ ¨ ( ¤ ( d Ñ 2 C Ð û ¤ æ ¤ " ( ¤ ( Now if is very small. then this integral is approximately . Thus ) ¨ E ) C ß f X ) ) ¨ ¨ Ú ß ° W ¤ ¤ Ñ ç E Ð 2 ¤ C f E E ¤ 2 C ( d 2 C E 2 C ð ¤ û ( 7 r g 2 for then E ) C ß E 2 C . ﬁnd a function The following theorem provides the answer.2: ð û û E E 2 ( § C ) ) C C f ¤ E ç E 7 ( ¤ E 2 ) C ¤ Ä Ä ( 2 ¨ q q C ( û " " C C ¤ 2 ( ( ð " C d û ( û ð Given a function . Take a sequence ç E 2 ¡ C ¤ d 2 C û æ Ï Î Í E 2 C ð û û Using differentiation by ﬁrst principles. THEOREM 9.8 THE FUNDAMENTAL THEOREM OF THE CALCULUS We are now in a position to be able to solve the following problem raised in Section 9. The theorem is proved using differentiation by ﬁrst principles.
8 Find the derivative of each of the functions in Exercises 1–8. We conclude that Similar calculations give . It is 7 ) ¨ " ° E 2 C û EXERCISES 9. Let us deﬁne a function domain by ) ¨ ) " ° 2 ö õ Í E l 7 % C This function is often denoted by .14) EXAMPLE 9. we found " ' ` % 2 ¨ " 2 ° 2 2 ö õ Í " 2 ) E 2 " ö õ Í C ) 2 ¨ ¨ 2 ó Í ) ~ " { ) ¨ ) " ¨ ) " ' ) " ` " ° % ' f f ° ` % ö g õ Í ö õ Í to 4 decimal places. 2 ¨ E § Ó d o 2 C ¦ ° ° ° E l E 7 ) C % C 2 ¨ Ê " ) f ) É o § ° ° ° E 2 r where is any positive constant and is any number in ) ß .188 INTEGRATION The function we are seeking follows directly from the fundamental theorem of the calculus. 4. The function value is equal to the area of the region under between the vertical lines and (Figure 9. ) ¨ ) ó Î ò Æ Å ° E ü C ¨ E " d ) o g C Ú ° E 2 C 5. It follows from the deﬁnition and the fundamental theorem of the calculus that (9.11 In Example 9.14) and is calculated by. ü ) ) 9. Simpson’s rule. ) ¨ E ) g ó Î ò d " C E 2 C ( ) ¨ E ) d " C E 2 C 1. for the curve example.8. 2. 6. 8. ) ¨ E ) C E 2 C ¨ E ü f " C ß E 2 C 7. C with . 3.9 THE LOGARITHM FUNCTION 2 ö õ Í The above example gives us a leadin to the logarithm function.
so we can write 2 ~ " E 2 C ð û 7 " ¤ d û In Section 9. { 2 ¤ d % " { ö õ Í " % E % C { ¤ % " ö õ Í 2 E 2 ¤ 2 ö C õ { Í 2 E 2 C { ö õ Í ¤ d 2 E E 2 C { C ö õ Í 2 ö õ Í E 2 C We now know that . that from inside the logarithm. we could write the (9. The solution to the initial value problem is . The only problem that remains is to write the unknown function explicitly in terms of . It follows from the deﬁnition of that .THE LOGARITHM FUNCTION 189 Figure 9. In order to do this.14: The logarithm function and " ³ § ° Let’s now return to the matter of solving the initial value problem { ¨ from which we found û To get the constant we use the initial value to get . to extract the properties of the logarithm function. we will need to derive some is. so that .15) i P ) " ¨ ) a g E r % 2 C g { " 2 E E 7 2 { r C 2 " 2 { ' ö C õ ö Í û 2 õ Í ¨ { " " ¨ .2 we showed that by choosing a function above differential equation as ç E 2 C { C û æ 2 ¨ ¨ such that . In the next chapter we will discuss the matter of writing explicitly in terms of .
It is not yet clear what happens as becomes arbitrarily large. this is what Mathematica essentially does. it follows from the deﬁnition of % % w u " 2 2 ö ö ö õ õ õ Í Í Í axis. for all in the domain of .15. the slope of the graph increases and approaches inﬁnity.15: Graph of " 2 P % f u E 2 E U 2 ö õ ö Í C õ Í C 2 2 ' ö ¨ 2 2 ¨ õ ¨ ö Í We can also ﬁnd the concavity of the graph of . In principle you could for different values. In fact. . As gets close to zero. or does it become arbitrarily large? A software package such as Mathematica has the logarithm function built in. Does the slope of the graph approach a number. 2 2 ö õ Í 2 ö õ Í 2 ¨ 2 2 % % w w 2 2 2 ö õ Í { Figure 9. In fact. plotting these points plot such a graph by using Simpson’s rule to calculate and then joining them with a smooth curve. that is. and we can use it to plot the graph in Figure 9. Since { õ 7 Í % 7 g " 2 w { E 2 E ö 2 " õ Í ö C õ Í 2 E C ¨ 2 l ¨ ¨ 7 % C " " g g % f f 2 ö õ Í Since the domain of is . Thus the graph is increasing and concave down for all . the graph does not cut the 7 " u 7 " 2 w u % 2 if if % " 2 The graph of crosses the axis at the point the logarithm in terms of integrals that 2 2 ö õ Í .190 INTEGRATION we have ¨ ¨ for all .
put r ¤ ¤ 2 ö õ Í " r " f ö ö ö ö õ 7 õ õ Í 2 Í õ Í % Í ö r d e ¦ õ f d ö Í d ¦ r õ r r r Í Since has derivative zero. If is a positive integer we have E f r § ¦ ¦ ¦ e d § r § r c C ö õ Í ¤ r ö õ Í terms f r ö õ Í d ¦ ¤ terms ¤ 2 . Here we will only prove the result if is an integer. Using the chain rule we have ö õ Í r f r ö r ö õ Í õ ö Í õ f Í i " r r r ö ö ö õ õ õ Í Í Í E 2 C ð . Then .3 is proved in a number of stages. so equation to get ö õ Í 2 ö õ Í r f ö õ 2 Í r ö õ Í r ö õ Í ¤ 2 ö õ Í f " 2 r ö õ 2 Í ¤ for some constant . put . To ﬁnd and rearrange the last % " 2 f r 2 " r 2 ö E õ Í 2 ö f õ d r Í ö ö ö 2 õ õ õ ö r r Í Í Í f õ ö ö Í 2 õ õ Í Í c r ö r õ ö Í E õ C Í 2 C 2 ¨ ¨ û To prove this. consider the function û .3 Properties of log For all positive numbers and and all rational numbers .3 now follows easily: " r The ﬁnal identity of Theorem 9. we have ö õ Í d r ö õ Í r ö õ Í È r Next we have r " ö õ Í r ö õ Í d " ö õ Í so that r ö õ Í f " r ö õ Í The next identity in Theorem 9. we have . Finally.THE LOGARITHM FUNCTION 191 THEOREM 9.
to be deﬁned. Thus (9. This is usually written 2 2 % l 2 ö õ 2 Í ö õ Í Ñ Ï Î Ð Í § 2 ' ö õ ` Í % ¤ g ö õ Í ' ` % g ö õ Í 7 ¤ r ö õ Í r ö õ Í We have shown that " 2 E 2 ö õ Í C 2 ¨ ¨ % u 2 7 ¤ d E 2 f C ö õ Í 2 ¨ " 2 ° Hence ü 2 ¨ ¨ E ü E ö " õ Í C f C ü ¨ " " ¨ ü 2 E E 2 f C ö õ Í C % 2 ¨ w ¨ ü 2 f ü % u 2 If . Then gets closer to zero. so gets arbitrarily large as increases. but we will not go into detail on this matter. Then % w 2 7 ¤ 2 d ö õ 2 Í ö õ Í 2 ¨ " 2 % ° w 2 In this equation. so that . We shall show on page 205 that this result holds for all rational numbers . Questions about the behaviour of for large and small values of can now be answered.9. We write (9. Since . if is a negative integer . we can write r ö õ Í .192 INTEGRATION We have shown that where is any nonzero integer.1 The indeﬁnite integral of 2 Ú % 2 The notation is used to emphasize that can only approach zero through positive values.17) . we have . this as . where is a positive integer. That the result is also true if is left as an exercise. we must have in order for 2 ~ " 9. It is also true if is any real number. so gets arbitrarily small as å r Ê " r å ö " r õ ¥ ¤ Í ö É r r õ Í ö ö ö f õ õ õ f Í Í Í Next. we put .16) 2 l f 2 ö õ Í É Ï 2 Î Ð Í ö õ Í § ' ` % f ¤ ¥ g ö õ Í Similarly.
We take logarithms to get E g d 2 C E " d 2 C Differentiating both sides of this equation gives " f " f " d " E 2 C ð Hence " f " f " d " Á E 2 C E 2 C ð E U Â U d " d 2 C 2 ö õ Í f U f '  Â d E U 2 "  d ' 2 ö d 2 d õ Í 2 2 f ' C ö g d õ { Figure 9. 2 .9.16.16: The graph of P U  ' 2  ö ' Í õ " d g Í 2 d f 2 E 2 g d { g " Â " d E d g U " 2 d 2 C d d 2 ö Á õ 2 ¨ Í 2 " " g g C " E E E d g f f U " ' d E " f d d " d 2 d 2 2 d 2 C C g C E E 2 Á 2 " ' f C E ö ö 2 d d C õ õ Í Í ' 2 2 C C f E  2 U C 2 f  ö ö õ õ r Í Í f { The graph of is shown in Figure 9. it may be easier to begin by ﬁnding the derivative of .THE LOGARITHM FUNCTION 193 We can combine equations (9.18) 9. The process is best illustrated by example. EXAMPLE 9. If we are given and want to ﬁnd .12 if Find E E g U d d 2 2 C C E E " ' d d 2 2 C C E 2 C E 2 C ð E 2 C ð E 2 C E 2 C ö õ Í Solution.17) into the single equation % k 2 7 ¤ d  2  ö õ Í 2 ¨ " ° (9.16) and (9.2 Logarithmic differentiation Logarithmic differentiation is a technique for simplifying the process of ﬁnding a derivative in certain cases.
three. § 2 d " C E 2 d " · 2 Ý C ö õ Í E 2 C 6. Suppose the eye is presented with a series of lamps of absolute brightness one. The next group of prominent. The stellar magnitude scale described in the previous example can be extended at both ends. The brightest stars were said to be of ﬁrst magnitude.9 Differentiate the functions given in Exercises 1–11. Similar conclusions apply to sounds. In fact. then g % f ' r r ¤ ó Í 13.512 times fainter than a sixth magnitude star. 2. E 2 2 f d " " ö õ Í E 2 C 8. eight. then it is the second series and not the ﬁrst which will appear to form a uniform gradation of brightness. In the nineteenth century. units. and so on.512 times brighter again. 5.194 INTEGRATION EXERCISES 9. it was found that a sixth magnitude star was almost exactly 100 times fainter than one of the ﬁrst magnitude. classiﬁed them in terms of magnitudes. It would only be visible in a telescope. and then with another series of lamps of brightness one.3 5 2. four. two. On this scale the sun is of magnitude and a star of magnitude . " E is . 12. down to the stars which are just visible to the naked eye. 9. E E E 2 C ó Í C ö õ Í E 2 C 4. What type of graph do you get? (c) Show that if is the brightness of a star. For physiological reasons. E " d 2 C ö õ Í E 2 C E U E ° d E ¥ " ¤ Ó 2 2 a 2 d d 2 C d ` d f d 2 g ö " g 2 2 " " õ C C C C f Í ö ö ö ö ö õ õ õ õ õ Í 2 Í Í Í Í E E E E E E 2 2 2 2 2 2 C C C C C C 1. all at the same distance. 7. In 200BC. but less bright stars were classiﬁed as being of the second magnitude. when accurate measurements became possible. four. 3. Magnitude Brightness 1 100 2 40 3 16 4 6. These considerations are the reason we adopt the stellar magnitude scale for measuring the brightness of stars and other celestial objects and the decibel scale for the measurement of sound. the Greek astromomer Hipparchus. A seventh magnitude star is 2. scales of luminosity and sound are often measured logarithmically. which were classiﬁed as being of the sixth magnitude. scaled to 100 for ﬁrst magnitude stars. Logarithms have a number of applications. and its magnitude. Similarly.512 times brighter than one of the ﬁrst is 2. a star of magnitude zero is 2.5 6 1 (a) By what factor does the brightness of a star change for a unit change in magnitude? (b) Plot a graph of the logarithm of the brightness against the magnitude. the following table of relative brightnesses was constructed. units. wishing to describe the brightness of stars. 11. magnitude " f E " f 2 C § E 2 g d " C ö õ Í E 2 C 10. two.
You should ﬁnd that relationship between the variables is approximately linear. Compute . such as elephants.5 3. (b) What do you notice about the plotted point for the hippopotomus? Suggest a reason for your observation. In land animals it is found that the leg diameter is (not surprisingly) related to the body mass. have relatively thicker legs than smaller animals. then ¤ ¤ ó Í ¤ has brightness and a star of magnitude has . Estimate the mass of the animal when it was alive. although some incomplete fossil skeletons suggest that larger animals may have existed. Here is a table giving the leg bone circumference (in centimeters) and body mass (in kilograms) for a selection of animals. (c) A fossil leg bone of a Brachiosaurus was found to have a circumference of 1. Mouse 0.05 Cat 4.6 0.THE LOGARITHM FUNCTION 195 (b) Let denote the brightness of the sun and let denote the brightness of the faintest . star visible to the naked eye.5 Human 13. What is the physical meaning of this ratio? determine 14. Use Mathematica to ﬁnd the best straight line through the points you have plotted. so that large animals.33 metres. The Brachiosaurus was one of the largest land animals of all time.4 350 Hippo 40 1700 Elephant 80. Leg thickness increases in proportion to the body mass rather than to the linear dimensions of the animal.5 6000 Ê E ¤ Ê ¤ ~ ¤ C ó Í Ê ¤ ~ ¤ ¤ Leg circumference Body mass (a) Plot the logarithm of the leg circumference on the vertical axis and the logarithm of the body mass on the horizontal axis. such as dogs.6 50 Horse 27. By trial and error with your calculator. E ¤ f C g % (a) Show that if a star of magnitude brightness .
a normal person can just detect a noise of 10 would occur at 100 W/m . Express this noise level in decibels. ¥ % " 9 Ê 9 É ó Í % " ó Í . Permanent ear damage square metre. As mentioned above. The noise caused by a jet aircraft takeoff at 60 m is 130 db. then the difference between and 9 ¥ 9 9 9 9 % " ¦ (a) Take W/m as the reference point on the decibel scale. (b) A pneumatic riveter has a noise level of 10 W/m . The range of human hearing is impressive.196 INTEGRATION 15. If is the intensity of a reference sound and is a in decibels is given by measured sound. the ear perceives sounds logarithmically and this has given rise to the decibel scale. If the intensity of sound is measured in watts per W/m . Compute the intensity of this sound in watts per square metre.
we have . . We want to solve problems of the following type: û + { % { ð { A function û ! is given.1 Suppose + û . {  + { ö õ Í . The remaining problem we have is to extract from this equation. Then .  +  + {   + { +  B +  + We let . Then if .CHAPTER 10 INVERSE FUNCTIONS In solving the initial value problem we obtained . Then gives  B + d  +  { C +  {  + { + û d C EXAMPLE 10. + { + We know ! ! How do we ﬁnd ? { and so û and so û E  +  + {   + { + 7 d  + We let . Let’s start with some simple cases which have a similar pattern to this one. This is done using inverse functions. Then gives 7 d 7 B   + +  { {  + { + û EXAMPLE 10.2 and suppose that Let 7 B + û .
Functions which have this property are called inverse functions . You should note carefully danger of confusing it with the reciprocal of in the notation for the inverse is not an exponent. using and from Example 10. The notation for the inverse function is universally used.198 INVERSE FUNCTIONS What is the relationship between the functions and in these examples? We can answer this question by applying the two functions in succession. If we think of a function as a set of operations which starts with an input number. A inverse of and is the inverse of .2 gives 7 B 7 d 7 d + û + + û In both these cases each function has the effect of reversing the action of the other.3 Let . Using and from Example 10. the other function reverses these operations and produces the number we started from. that is. with . is the instead of for the inverse of . We normally write function and its inverse are related by the pair of equations û ¥ û % + ¥ û + û ¥ û û û û This is illustrated in terms of function machines in Figure 10. Find the inverse Y 7 d + Y + ¥ ¥ û + û + û ~ + û + û ~ + ¥ û B + û ~ of . The symbol does not mean that the . We will postpone a more precise deﬁnition of inverse functions to the next section.1 we have d B û + § C û B + + û + + û These relationships are illustrated in the diagram below. operates on it and produces an output number then.1: The action of a function and its inverse û Similarly. Figure 10. Unfortunately. in the above examples. EXAMPLE 10.2. there is the . If we do want to express in exponent form we will write . Y 2 2  Í Ë 7 B  d 7 d B 5 C + Ì Ì + µ Ï 7 + 2 2 3 B û Î + · ¹ + ¥ d û C + ¥ + û Ë Í + + û + 2 2 .
as before.199 Figure 10. we let . If we know that a function has an inverse and if 7 2 E E 2 C { C û ¥ û û i 7 § 2 E { C E 2 C ¥ { Alternatively. Then P P · Ð Í Í § S P S Q P · Q Ð Í Í · Ð Í Í P P . We put 7 2 ç E 2 C ¥ æ so that 2 U d E 2 C ¥ ' Hence E U f 2 C E 2 C ¥ so that 2 U d { ' Hence E U f 2 C { and. E U f 2 C § E 2 C ¥ Now we can solve the problem posed at the beginning of this chapter.2: The action of a function and its inverse Solution.
There is a graphical interpretation of 1–1 functions. (Will this never end?) How can we be sure that a given function has an inverse? ! How can we calculate function values for the inverse.1 THE EXISTENCE OF INVERSES Consider the function deﬁned by . assuming it exists? ! 10. Any function has the property that an input number gives only one output number. which is impossible. if . because functions must give a unique require output number for each input number. We write 1–1 rather than onetoone. In Figure 10. You are asked to show in the Exercises that a function which is either increasing on an interval or decreasing on an interval will be 1–1 on . but we now have two more problems to deal with. then . This would and . if . On the other hand. so that any vertical line will intersect the graph of the function in at most one point. the graph on the right is not 1–1. This test cannot be applied if the function is not differentiable at one or more points and other methods are needed. if we take the function . then implies that . If were to have an inverse . Hence in this case there can be no such function and the function does not have an inverse. { + Y ¥ { C B C { ® C B + + + ¥ 7 { C C + + ¥ d C C + B Y 7 + ¥ Y C 7 ¥ + ¥ + Y  + ¥ û  + { û ¥ û Hence.3. while the graph on the left is 1–1. A function with this property is said to be onetoone . gives us . since undoes the action of . then there is only one input number which will give this output number. Hence if the derivative of a function has constant sign on an interval . For example. For example.  + ¥ û    + { + û + ¥ û ¥ û then. A 1–1 function has the additional property that any horizontal line intersects the graph at most once. These arguments can be generalised to show that a function can have an inverse only if each output number comes from a unique input number. If we pick any output number and put . then there are two input numbers for each nonzero output number. applying to both sides of this equation.200 INVERSE FUNCTIONS The initial value problem + { % { ð { thus has solution + ¥ ö õ Í + { This is all very well. then this inverse would satisfy and . then the function will be 1–1 on .
by the remark on page 200.3: (a) A 1–1 function. If is in the range of . there can be only one such input number and we deﬁne to be this number. then it must come from some input number in the domain of . We equate and and if this forces us to conclude that . so that. 2 ¨ 2 ¨ Y (a) (b) ¥ Y . Then there is a unique function range of and which satisﬁes the equation + + Y Y Y whose domain is the The proof is straightforward. We leave it as an exercise to show that the inverse of is given by E Y B + Y + ° + + Y ¥ Y d 7 Y M 7 d 7 + ð Y Y d 7 + + + Y Y Y + + Y Y Y for all in the range of .1 Let be a 1–1 function. We are almost ready to give a formal deﬁnition of an inverse function. Choose any two real We show that the function numbers and . then is 1–1. ¥ Y DEFINITION 10. The function we have called in the theorem is called the inverse of and is usually denoted by . This means that if . The inverse of . The function is 1–1 if distinct input numbers never give the same output number. THEOREM 10. (b) A function which is not 1–1.  + Y Y Y Y for all in the range of . First.1 Let be a 1–1 function. is 1–1. Thus the The equation function is 1–1. is the unique function EXAMPLE 10.THE EXISTENCE OF INVERSES 201 Figure 10. . then necessarily . implies that and hence that . denoted whose domain is the range of and which satisﬁes the equation Y + ¥ Y + Y Y Y . we state the following result. Alternatively. Since is 1–1. we see that .4 is 1–1 and hence has an inverse.
normally written and the inverse of is . The graphs of these functions are shown in Figure 10. r Ä j r r r is the unique  + ¥ Y Y Y Figure 10. We denote the th root of by or . DEFINITION 10. the function and its inverse enable us to give a precise deﬁnition of square roots. We denote the r r In a similar way. let be any positive number. The fact that inverses are unique guaran.2 The square root . In fact. the function h g % + û does not have an inverse. The nth root of positive number such that . but where it would nevertheless be very convenient if it did. Then the square root of is the unique positive number such that Let square root of by . For example.  + 2 ¥ { C C C B B B + { 2 + + ¥ Y Y { { C B + ¥ + C C ¥ B B Y B .4: Graphs of (a) and its inverse (b) and its inverse. but both the functions % + Y and do have unique inverses. Often we can deal with this situation by restricting the domain of the original function.4. The inverse of is tees that has a unique positive square root. Then ¥ Y Y + ¥ Y + Y Q % + so that ç + ¥ Y æ It is the number which is called the square root of . To do this. we can deﬁne th roots of any positive number. ¤ r ¤ r r r .202 INVERSE FUNCTIONS (a) ¨ (b) ¨ There are cases where a function does not have an inverse.
we get of all points of the form as the output number.1 The graphs of and . . Figure 10.5: The slopes of and P E 2 7 { C Ñ 7 { E ~ ~ ~ ~ 2 ó ó ó ó C Ñ E 2 . Hence the graph of consists of all points of the form .4 shows this symmetry in two particular cases discussed earlier. Notice also that the range of will be the domain of and the range of will be the domain of . The graph of consists of the set . Since the . 2 ¥ E E 2 C C E ¥ 2 7 E 2 C C ¥ 2 ¥ { E 2 C ¥ E E ¨ 2 C 7 2 C ¥ The relationship between the slopes of the graphs of and is also a simple one. Thus ~ " ¥ Figure 10.THE EXISTENCE OF INVERSES 203 There is an important connection between the graphs of and .1. This means that the two graphs will be symmetric with respect to the line . it follows that the tangent line through graphs are mirror images of each other in the line on the graph of is the mirror image of the tangent line through on the graph of . If we take as the input number for . ¥ ~ E { ~ f f d d 7 2 C " " " " 2 C { { ¥ { 2 10.5). . E E E 2 C ¥ { 7 7 2 2 C C while E ~ f U ~ C ó . Then ¥ E E 2 7 2 E 7 2 C { C C ¥ E ~ 2 d U ~ { C ó . . With the aid of trigonometry. it can be shown that the slopes of these two tangent lines be the slope of the tangent line to the graph of at the point are reciprocals of each other. Let and let be the slope of the tangent line to the graph of at the point (Figure 10.
1) 2 " g E 2 C 2 ¨ ¨ The reasoning we have applied to square roots can be applied to th roots for any positive integer . We have E 2 E C 2 ð C E ¥ ¥ C 2 ~ " E E 2 C ¥ ¥ 2 g C ¨ E 2 C ð E 2 C E ¨ { % 7 w 2 C 2 E 2 7 { C 2 E 2 C 2 so we have shown that (10. You will need to be thoroughly familiar with function notation and you will probably have to think carefully about it before it begins to make sense.3) .3 The derivative of For all rational numbers . E 2 C THEOREM 10. Our main use of Theorem 10.2) can be used to prove the following important result. % u 2 ¤ 2 ¥ U ¤ j T 2 " E ¤ j 2 C 2 ¨ ¨ % w 2 If THEOREM 10. The theorem states that if the slope of to the graph of is .2) (10. This is a rather difﬁcult the tangent at the point theorem to come to grips with. We can show that for is odd.204 INVERSE FUNCTIONS where . then is differentiable at and " ¥ 2 ð ¥ ¥ % k ç E 2 C ¥ æ ð E E 7 E E E E 2 2 C C 2 { E C " C ð ¥ " ð 2 7 ¥ E ¥ i " 2 2 C C C ¥ " ð ð ð 2 g ¨ C E E 2 i C E ð 2 E 2 C E C 2 { E ¨ ¥ C ¨ ð 2 C ¨ C E 2 C ð E E 2 ¥ C C ¥ Since and .5. The chain rule. we have { is differentiable at and if (10. The inverse function is given by Let . The value of is the slope of the tangent at to the graph of .2 will be to derive formulas for the derivatives of various functions . of interest.2 The derivative of an inverse function . we will use it to calculate the derivative of with domain . If Let be a function having an inverse function . together with equation (10. ¥ You should try to understand this theorem with the aid of Figure 10. we have ¤ 2 The result holds on every open interval on which is deﬁned. then . then this formula also holds for . Then . As an illustration. This relationship between the derivatives of a function and its inverse is summarised in the next theorem.
but was only proved for integer values of . Use the chain rule and equation (10. This result was ﬁrst stated in Theorem 9. " 2 d 2 E 2 C 2 d § 2 E 2 C 3. . 14. § E 2 f " C E 2 C " f § 2 E 2 C 9. EXAMPLE 10. The idea is to solve the equation for . Putting ¤ shows that .CALCULATING FUNCTION VALUES FOR INVERSES 205 We conclude this section by showing that ¤ 2 ö 2 C 2 7 õ ö ¥ E Í % " õ Í 2 { C d ö " õ ¤ 2 Í f 2 ö % ¤ õ Í 2 w f for all rational numbers . We have ¥ ¤ 2 ¤ ¨ Hence ¤ EXERCISES 10. Show that the function .3 of Chapter 9.2 to derive equation (10. 8. 2 2 d " E 2 C 11. 4. . Hence 2 f 2 7 ¤ ¤ f 2 2 % ö õ Í 2 ö õ Í C 2 ¨ . " f 2 E 2 C ' d 2 g E 2 C 1. It is easier to work with the symbol rather than the string . 10. " ' f d 2 2 E 2 C " d " § 2 E 2 C 5.3).1 Determine whether or not the given function is onetoone and. has a unique inverse together with Theorem 10. ﬁnd the inverse function. Use this result 2 ö õ Í 2 ö õ Í " 2 where is a constant. Let 7 r be a function which is either increasing on the interval . 10. 6. or decreasing on the interval .2 CALCULATING FUNCTION VALUES FOR INVERSES In simple cases we can use algebraic manipulation to ﬁnd an exact expression for the inverse of a given function. where possible.2) to derive equation (10.2). 2 ¤ 2 E 2 C 7 r 7 r 12. Show that is 1–1 on . You may prefer to write for .5 Show that the function % w 2 E 2 C E 2 C ¥ E 2 C ¥ 2 E E E 2 C ¥ { ¤ j 2 E 2 C ¥ 13. g d 2 r E 2 C  2  E 2 C 7. 2.
this is not usually the case.206 INVERSE FUNCTIONS so that d d In examples such as the one above we were able to ﬁnd a formula for the inverse function without a lot of difﬁculty. It follows that and solve . The equation of the tangent at is B + + + ð + Y Y Y + + Y Y + { + Y Y % B { + + Y This intersects the axis at + + ð Y Y B and will (usually) be a better approximation to than .6 shows the graph of . We can usually arrive at an approximate Suppose we want to solve the equation solution by various means—plot the graph of and estimate the root or plug in values of into until we get one that gives . In general. Figure 10. We repeat this process by taking as the new starting value. However. Suppose + Y . Then we have is 1–1. Let denote the exact solution we are the formula for seeking. . Newton’s technique was modiﬁed by Joseph Raphson in 1690 and by Thomas Simpson (of Simpson’s rule fame) in 1740. We get % d d  + Y ¥ Y is 1–1 and ﬁnd its inverse Solution. if we have an approximation to the root. . for example.2.1 Newton’s method + ¥ Y The idea behind the method of ﬁnding roots of equations known as Newton’s method was discovered by Sir Isaac Newton around 1660. How. the essential idea is Newton’s. can we ﬁnd the inverse of more complicated functions such as the logarithm function? We need a method of calculating values and one way of doing this is to use Newton’s method. Nevertheless. The form in which it is used today is not precisely the one discovered by Newton. . since both of them are positive. The method is often called the Newton–Raphson method. let . then the next will be approximation ¤ ¤ ¤ + + ð Y Y B ¤ Ú ¤ Ú ¤ E B Ý and hence + ¥ Y % B Ý so that { Y % d  { + { Y  + ¥ Y { % ¥ Y Hence To ﬁnd and therefore . of 10.
. Then g ¤ f 2 ' § ¤ g 2 ¤ d 2 f ¤ § ' 2 ¤ g 2 2 EXAMPLE 10. Suppose we for some number in the domain of . g g ' ' ` % r r ` g g g g " " " " g g g g § d d d d 2 2 2 2 § § § § § ' ' ' ' 2 2 2 2 g g g g § ° 2 2 2 2 r " 2 Take as a ﬁrst guess. Newton’s rule for this case is Ú ¤ 2 % g f § 2 will be a cube root of 2.6: Newton’s method We thus generate a sequence of values Our main interest in Newton’s method will be for calculating inverse function values.CALCULATING FUNCTION VALUES FOR INVERSES 207 Figure 10. which in most cases converges to the root ¡ P E 2 C 2 { 2 2 ¤ 2 ¨ .6 Let’s solve the equation . Since and 2. are given a function and that we wish to calculate We have ¥ r E r C r ¥ E E r C ¥ C i g r g " 2 The convergence is rapid and we get . we expect it to lie between 1 2 .
2 % 10. we note that " since the domain of is . so the above equation can be written 7 % w 2 E 2 ! g ' g ! ` % E g C ¥ We conclude that . which function has an inverse. we have to solve . Then % w E 2 C ð EXAMPLE 10. we have to solve the equation that is. Notice that is welldeﬁned since Let for all . By Theorem 10.4) ¥ E E 2 C " ¥ C ð 2 ö 2 õ ¥ 2 C C E l 7 % C 2 ö õ Í E 2 C E 2 " ¥ C ~ ¥ " E C 2 ~ " E 2 C ð Now . a ! ! g g U U % ' ' ' r g g g ` g ! ! ! ! a ` ` ` ` % % % % % " " " " " " " " " " d d d d d d d d d d § ³ ° 2 2 2 2 2 " 2 Take as a ﬁrst guess. Hence is an increasing function on its domain. Newton’s rule for this case is g f E 2 C ¥ " d § ¤ 2 E g g C g f " d 2 % d w § 2 2 " d 2 d § 2 E 2 2 C E r C ¥ 2 " d ¤ % 2 ' r E r C ¥ 2 Hence putting and solving for gives us . This shows that the notation of Theorem 10. . In order to conform with the implies that it is 1–1.2 we have (10.208 INVERSE FUNCTIONS so that % r f E E r C ¥ ö Í E ö 2 ¥ C § § § § § § § õ ° ° f 2 õ ö Í ð Í C Ú 2 2 2 2 2 2 2 2 2 2 E õ E % C ð Í ¤ g g g g g ' ' ' ' ' E 2 C 2 C 2 ¨ ¥ E ¨ for . . Determine . Firstly.2 The inverse of the logarithm function We can now apply our knowledge of inverse functions to the logarithm function.7 .2. Then the inverse of the logarithm function is .2. let us put . According to the discussion in the text.
5) have one and only one solution. We can now calculate it in a totally different way.4 We have 2 Ô Ó Ò Ó The relationship between the logarithmic and exponential functions is shown graphically in Figure 10.3 Computing the exponential function 2 Ó We have already computed for various values of using inﬁnite series. Since the exponential function is the inverse of the logarithmic function. In other words 2 ¥ ö õ Í E 2 ¥ ö õ Í C ¨ whose solution we know to be the exponential function. problem . Using Figure 10. The fact that exponentials and logarithms are inverses of each other proves the following theorem.7: as the inverse of Ó { ö õ Í Ú ¤ { % w O 2 for all . THEOREM 10.5) Ó E 2 C ¥ ö õ Í This is in fact true. Recall (page 207) that we can compute inverse functions by using Newton’s method.7.CALCULATING FUNCTION VALUES FOR INVERSES 209 Hence the inverse of the log function is its own derivative. We would like to conclude that (10.2. we can compute by solving the equation Let’s try computing .7. Hence satisﬁes the initial value . we Ó 2 ö õ Í % ~ " 2 f f E ¤ Ó { C Ó Figure 10. so that 2 ¨ . but it is important to realise that we are using a result from the theory of differential equations which states that initial value problems such as (10. 3 2 2 ö õ Í { 1 1 2 10. 4 3 4 E 2 C ¥ ö õ Í 7 " E 2 C E { % C { 7 E 2 C " { E % 2 C ð ¥ { ö õ Í % " ö õ Í Furthermore. so that we have to solve ¤ { ö õ Í " for . Then using guess N M { g { { % " 7 f " { ö f ¤ õ { Í .
then the values have to be calculated by a method such as Simpson’s rule using the deﬁnition of the logarithm as an integral.210 INVERSE FUNCTIONS we get " a ' " ` g " { f " " " { ¥ Ú d ö Ó Ó g 7 f f f § õ " " { Ó £ 2 Í 7 ö { { { { { § E õ ~ ~ ~ ~ { Í { { { 2 " " " " C ö ö ö ö E E E { { E { õ õ õ õ { % { Í f C õ ¤ 7 7 Í Í Í Í 2 C ö 2 C C ö õ ð { Í f f f õ { { E Í C § 2 C { { { { 2 ð ¨ ¨ { § ° { Thus . Ó . we ﬁnd ! ! U g g g ` ! ! £ " " " a a a g g g { { { £ . Thus E ¤ d 2 C ¤ where .2. If we are to be totally honest in the above calculation and not use builtin functions . and will in calculators and/or computers. the solution to . We have so that Hence 7 for some constant .4 The air pressure problem ! g ! " a g ö õ Í U g ` " a g ö õ Í " a ' " ` g ö Ó ! g ! " a g Ó In Chapter 6 we solved the initial value problem { E % by using inﬁnite series. it is a consequence of the theory we have developed. but produce exactly the same result. Rather. 10. we can now solve it by a quite different method. . Consequently. This is not an obvious fact. Truly a heroic calculation! The calculations of by power series and by Simpson’s rule and inverse functions are completely different. Using the material developed in this chapter. Using the initial condition the initial value problem above is as we have previously shown.
0002% of the brightness of sunlight. In clear seawater. We have to ﬁnd both . Ó Õ ¤ Ê " d " É Ô Ó converges to . sequence Õ THEOREM 10.1% of the surface value. and from the given information and then use these values to to get .CALCULATING FUNCTION VALUES FOR INVERSES 211 EXAMPLE 10. then Ó r g % r g % " ( Next. where represents the distance (in metres) travelled by the light. What is the light intensity 5 m below the surface? Solution. but it only indicates that the human eye is able to function effectively over a vast range in light intensity. This equation can be integrated to give .5 The sequence converges to . it is found that the intensity of light 1 m below the surface is 25% of the intensity at the surface. so that 7 ¤ 7 q d Ó ( ± £ E E ( ( C C ö õ Í £ ± £ where is a constant. the rate at which the intensity decreases is proportional to . the sequence i or about 0. First we prove a theorem on the convergence of the .8 When light passes vertically through a transparent substance. if . the brightness of moonlight from a full moon is (on average) 0. we put . Firstly. Finally. In particular. 1 This value may seem to run counter to what we expect from experience. put . to get q Ó £ ± % ( Ó E r where ﬁnd C . We can write E ( C 7 E ( C ± ( ¨ ¨ ( E ( C We conclude this section by proving two results which make use of the inverse function relationships between exponentials and logarithms. For example.1 ³ E r Ö g 7 ¤ ³ % ÿ E C " 2 ³ % Ó § C Ó % d % " þ r ( ¡ ¤ E ~ 2 r g d % " C Ó Hence .
4. We write E 2 C E 2 C ð r We call the general exponential function. If is any rational number. 2 r Equation (10. the limit of the term " ö õ Í f E ~ 2 d " C is just Thus 2 ¤ 2 d " " E " C ð ö õ Í so that Ó ¤ 2 d " The second result is the deﬁnition of irrational powers of a real number.7) Â Ø " " ö ö õ Í õ Í f f 7 E ÿ Ó ~ ÿ ~ 2 ö 2 2 õ Í d d " " þ C þ ö ö ~ ÿ ÿ õ õ Í Í 2 × Á 2 (10. rational or irrational.6) holds for all rational numbers .212 INVERSE FUNCTIONS We have 2 d " ö õ Í ¤ 2 d E þ þ ö 2 2 i C C õ ö ð Í ö r ÿ õ õ Í Í l " þ ö õ Í The thing to notice is that. then by Theorem 10. as . Rational powers such as and can be deﬁned using ordinary arithmetic. . but this is not the case for irrational powers such . by (10. We now deﬁne Ô Ó Ò ß Ó r for all .9 Find if Solution. we see that as i Ô Ó Ò ß Ó Ú ß Ô Ó Ò Ó ° j § g § Ù g § g DIfferentiating both sides gives 2 ö õ Í d " d 2 E 2 2 ö ö E õ õ Í Í Ú E " d 2 C d " 2 ö 2 õ 2 C d Í " E 2 C Ú 2 EXAMPLE 10.6) .
17. A bacteria culture increases by a factor of 6 in 10 hours. their method of ﬁnding as a better guess for . ' 7 " 7 % r f 2 g f 2 g 7 % 7 2 f 2 ó Î ò 16. 22. Show that the Babylonian method in the above example is equivalent to using Newton’s method . g 7 " 7 % g f § 2 g 7 " 7 % g f 2 12.CALCULATING FUNCTION VALUES FOR INVERSES 213 Consequently. ' 7 " 7 " d 2 U f § 2 g 7 % 7 % 2 ò õ ô f 14. apply Newton’s method to ﬁnd the value of " g ) ¨ ) ) d " 2 18. Ê Ê 2 2 ö ö õ õ Í Í d d " 2 2 d d " " É É E Ú 2 C 2 E 2 C ð 3. Use this method to evaluate to 4 decimal places. r E r C ¥ In Exercises 19–22. 24. This process was repeated until the to take the average of and desired accuracy was reached. Ù g E 2 C E 2 ò õ ô C E 2 C 9. Apply Newton’s method in Exercises 12–17 to ﬁnd a root of the given equation in the indicated interval. 13. In was to make an initial guess for and then modern notation. The ancient Babylonians discovered a way of extracting square roots about 4000 years ago. 15. on the equation 25. 8. E Ó 2 C ö õ Í Ô Ó Ò ¥ Ó Ó Ò ¥ 4. Ù E 2 ö õ Í C E 2 C 11. Use Newton’s method to ﬁnd a value of ° for which at the given point . 2 23. Ó 2.2 Simplify the following: Ô Ó Ò Ó E Ó C ö õ Í Ô Ó Ò Ó 1. 5. 6. . g r 7 Ó E 2 C g r 7 § 2 E 2 C 19. " EXERCISES 10. Ô Differentiate the following functions: Ý Ü Û 2 E 2 C 2 E 2 C 7. 10. 20. How long did it take the number of bacteria to double? % r f 2 r 2 ` r r 2 ~ r 2 r % r 7 " 7 % g 2 7 2 ó Î ò E 2 C ' r 7 2 ó Í E 2 C 21.
the C decays and is not replaced. y y y y y y ° ° ° ° ° ° ° 10. At that time the population was increasing at a rate of 380 000 people per day. a sample of a radioactive element Ra (a) How long does it take for the sample to decay to half of its original value? This is called the halflife of the element. determine when the coffee will have cooled to 50 C. The body is speedily taken to the morgue where the temperature is kept at 4 C. Estimate the time of death. 29. 30.8) " E % C ð { 7 % E % C { 7 % { d ð ð { by means of inﬁnite series. we ﬁnd ð { ð ð { ð ð { ð { g § ç E ð { C æ ¨ and that then the above equation can be written % ç { d E ð { C æ ¨ 7 ð { { ð { { g § ç { æ ) ¨ ¨ y y ) ¨ ) ¨ .214 INVERSE FUNCTIONS 26. After one hour. A body at a temperature of 32 C is discovered at midnight in a room of temperature 25 C. This gives % ð { { ð { d ð { ð ð { Using the chain rule. When would the world’s population be expected to double? To triple? 27. because the C is continually replenished by the action of cosmic rays on nitrogen in the upper atmosphere. (b) How long does it take for the sample to decay to 20% of its original value? Atmospheric air contains carbon.3 THE OSCILLATION PROBLEM AGAIN In Chapter 7 we solved the differential equation (10. We begin by expressing the left hand side of equation (10. The trick in doing this is to multiply the equation by . Radioactive elements decay at a rate proportional to the amount of the element present. Living organisms absorb both the C and the C and the ratios of these substances in a living organism is the same as in the atmosphere. Suppose decays to 58% of its original amount after 3 days. However. We will now solve it using some of the ideas in this and the previous chapter. In 1987. Assume that the temperature of a cup of coffee obeys Newton’s law of cooling. 28. but there is a small percentage of radioactive carbon (C ). almost all of it in the form of carbon dioxide. This radioactive carbon decays with a half life of 5600 years. Once the organism dies. the world’s population was 5 billion. Determine the age of the bone. Suppose the coffee has a temperature of 95 C when freshly poured and has cooled to 90 C one minute later. Most of the atmospheric carbon occurs as C . This is the basis of carbon dating. so that the ratio of C to C changes with time and this ratio can give an indication of the time since death. A bone excavated from an archeological site is found to contain 1% of the original amount of C . the temperature of the body has fallen to 25 C. If the room temperature is 25 C. the ratio of C to C in the atmosphere remains approximately constant.8) as a ﬁrst order derivative.
but the total energy is constant. for some constant . This statement assumes that there is no friction in the motion. the above condition becomes ¤ ± This says that the quantity on the left hand side is constant throughout the motion. but this turns out not to be possible. Hence which we write as (10. The left hand side of equation (10.11) û " ) { ¨ ¨ { f " " · { 7 { f " · ð { { f " · ® ð { Following our method in the last chapter. so the solution is E % C û d ) r E E ) C { C û r d ) E E ) C { C û E % C û r allows us to ﬁnd % w E % C ð ¤ { % f { " { ± { ± · d d d ð ð E E { ð ð { { C C ¤ (10.10) is the total energy of the particle. which is disguised by the way it has been written. If we take the differential equation for the particle of mass on a spring with spring constant . energy is changing from one form to the other. we use the initial conditions again. and in equation (10.9). During motion.10) . Since negative sign. We deﬁne û û ü ¨ ü f " " § ° E 2 C û û % E % C { and hence . we deﬁne a function " E { C ð û whose derivative satisﬁes Then equation (10. The initial condition .THE OSCILLATION PROBLEM AGAIN 215 and so (10. We might expect to ﬁnd an algebraic function for . The constant is called the energy and so we can say that the energy is conserved in the motion. The ﬁrst term is the kinetic energy of the particle while the second is its potential energy . so we to use the device we developed in the previous chapter and deﬁne in terms of an integral. we ﬁnd and so By using the initial conditions ¤ " " E % { C ð { d E ð { C % E % C { This gives us two possibilities: To decide which of these to take. we discard the . This last equation has a physical interpretation.11) becomes " ÿ E E ) C { C û þ ) ¨ ¨ All that remains is to ﬁnd .9) 7 where is a constant.
For ease of notation let . We can also show directly from the deﬁnition of and the use of Theorem 10. by the chain rule E ) C ð 9 E ) C 9 ð ð " E % C ð 9 7 % E % C 9 7 % 9 9 Putting all this together. the derivative of the inverse is " û E ) C ð 9 Then.2 that satisﬁes the initial value problem (10. As an exercise. . you should show that is 1–1.8) is . If above solution becomes 7 ) E E ) C { C û 2 . that . A judicious choice simpliﬁes things. The derivative of is given by an inverse E ) C û ) û ó Î ò ¥ û ) ) ó ) Î ò E E E ) ) C C û ) û ¥ ¥ ó Î û ò û C û E ) C E % ) C ¥ û ¥ " E % û û C ¥ ) 9 û ó Î E ò ) C ð ¥ û û ¥ û % ) ó E Î ò % C û g ~ 7 g ¥ ~ E û ) û C f E ¥ ) C û ¥ û ó Î ò From Theorem 10. and so equation (10. is that a suitable restriction of will be 1–1 and it is this restricted function that we will have to identify . " 2 " E E ) ) f C C E ) 9 9 ) C 7 E f " 9 E f f ü ) " E ¨ C " " ) f C 9 C · · " 9 ü ð ~ f f " f · " " E ) C ° with E 2 C û % E % C û % 2 Here can be any constant. has vertical asymptotes at We already know from Chapter 7 that the solution of the initial value problem (10. we should not be too hasty in concluding that .2. Consequently. However.8). we see that d ð ð 9 satiﬁes the initial value problem E ) C ð " 9 E % C 9 f " · E % C ð 9 % E % C 9 % E % C û Also. A function and its inverse are both 1–1 functions and certainly is not a 1–1 function.12) as E ) C { ) E ) ó Î ò C û " ® " ü 2 " f ) ó Î ò E ) C { A function which has an inverse satisﬁes the equation for all in the range of .12) (10. What is true though.13) suggests that we write . or equivalently. then and the (10. has . so that we can write equation (10. we must have and so it is the restriction of to the with interval which has to be identiﬁed with . the function being integrated) . so and from the above equation for .13) 9 .216 INVERSE FUNCTIONS The restriction is necessary because the integrand (that is. Since .
8: A Mathematica plot of the graph of . j § E ) ) f " C E ) C ð ð û û is always positive and so is an increasing function.5 % ) and hence there is a point of inﬂexion at . We have shown that the exact values are from the way we have deﬁned .1}.0.0].{x.1. To get a nicer looking plot.5}}] The graph is shown in Figure 10. The function is the inverse of and is called the inverse sine function..5. PlotStyle>RGBColor[1. 1. to the domain of .8. The second derivative is û ` ) " f " f " % E " E % f E C C ) C û ð û The plot suggests that and and .1}.0. This terminology is not strictly domain is not the inverse of .THE OSCILLATION PROBLEM AGAIN 217 Hence. this notation is well established and the alternative is too clumsy in use. try the following Plot[Evaluate[Integrate[1/Sqrt[1tˆ2]. The basic command is Plot[Evaluate[Integrate[1/Sqrt[1tˆ2]. it is the inverse of the restriction of to the domain correct: .{t. Also. The function has and range and is denoted by .5 1 1.1.5 1 0.{x. Ticks>{{1.{1.1.1.5.0. by the û Figure 10. û ` " g ~ E " C f û E " f C û g ~ E " C û .1. However.5.x}]].AspectRatio>Automatic.{t. Clearly Fundamental Theorem of the Calculus..x}]]. as before. What more can we say about the properties of the function ? We begin by plotting its graph using Mathematica.5 1 1 0.1}] û " 7 " f 2 û ) ó Î ò 2 û ¥ ó Î ò ) E ) ó C Î g ò ~ 9 7 g ~ f E ) C û " E 7 ) C " g ~ f û 7 g ~ f g ) g f 7 ) ó Î ò E ) C 9 E ) C ¥ û We have restricted instruction.
The inverse of this restricted can restrict the function function is known as the inverse sine function . The inverse function relations are .4 INVERSE TRIGONOMETRIC FUNCTIONS In the previous section we met the inverse sine function.9: (a) A restricted domain for .9(b). Once again. We denote E 2 ¥ ó Î ò C ó Î ò by 2 ó Î ò Figure 10. interval on which it is 1–1. The deﬁnition of the inverse sine function is given in the box below. Figure 10. In this section we will discuss some of the properties of this and other inverse trigonometric functions. . we have to restrict the function to an .218 INVERSE FUNCTIONS 1 10.3 The inverse sine function The inverse of the function g 2 g f 7 2 ó Î ò E 2 C 2 ¥ ó Î ò 2 ó Î ò We can also deﬁne the inverse cosine function.4 The inverse cosine function The inverse of the function ¥ ¥ 7 % 2 ò õ ô Figure 10.9(a) shows the way in which we to a part of its domain on which it is 1–1. The conventional choice is DEFINITION 10. The graph of is shown in Figure 10. We denote are 2 E 2 ¥ ò õ ô C ò õ ô by g " 2 2 ¥ 2 ó Î 2 ò g " f f % E 2 C 7 7 7 2 2 2 ¥ 2 ò is called the inverse sine function. (b) The inverse sine function. DEFINITION 10. The inverse function relations to a part of its domain on which P " ¨ f " f P g õ E ô E 2 2 ò õ ó Î E ô C ò 2 C C ¥ ¥ ò õ ó Î ô ò " ¨ f f (a) (b) g f .10(a) shows the way in which we restrict the function 2 " ò õ 2 ô 2 " E 2 f % C 7 7 is called the inverse cosine function.
4.14) P " ¨ " f P ò f " C C ð " ò " g ó õ Î Þ ò ô " ¨ (a) 2 ¥ ò õ f ô it is 1–1.INVERSE TRIGONOMETRIC FUNCTIONS 219 1 10. (b) The inverse cosine function.14) must be negative and hence 2 ¥ ò õ ô { ò õ ô f " · ® { ó Î ò " { ó Î ò d { 7 ò { õ ô ò õ ô f " { ó Î ò { ò õ ô { ó Î ò We can write in terms of by using the identity .2. let E 2 ¥ ó Î E ò E C 2 2 " ó 2 Î ¥ ò ò ó 2 ó " Î õ Î ò f ô Figure 10. The graph of is shown in Figure 10. there are two ways to proceed. For the inverse sine function we have ¥ " E 2 C ð E ¥ ó Î ò C so that To decide which sign to take. We have 2 ¥ ò õ ô { { ó Î ò f E { ò õ ô C { ¨ ¨ 2 ¥ ò õ ô To ﬁnd the derivative of and { ò õ ô .10: (a) A restricted domain for . we have to use the fact that the slope of the graph of in Figure 10.10(b) is negative.10(b). Then (10. (b) f g f 2 . .1 The derivatives of the inverse trigonometric functions The derivatives of the inverse trigonometric functions can be found using Theorem 10. so that the left hand side of equation (10. Firstly.
g ~ g so that 2 ¥ ò õ ô d 2 ¥ ó Î ò Ê 2 f " g 7 ¤ g 2 2 2 f ¥ " ¥ ò ò õ % õ ô ô d d 2 2 ¥ ¥ ó ó Î Î ò ò 2 2 ¨ ò õ ô d 2 ó Î ò show that is a constant. Then 7 % ¥ ò õ ô d % ¥ ó Î ò ¤ % This result is the reason for the coterminology for trigonometric functions. We take { . Put 2 ¤ . The inverse of a function and its cofunction sum to . We have " f É d " E 2 ¥ ò õ ô d 2 ¥ ó Î ò C ¨ Hence where is a constant.2 directly: " E 2 C ð E ¥ ò õ ô C EXAMPLE 10.10 By differentiating ¥ ¥ Solution. we can use Theorem 10.220 INVERSE FUNCTIONS Since 7 { 2 ¨ ¨ ß " { 2 ¨ ¨ we have 2 f " " f E 2 ¥ ò õ ô C 2 ¨ ¨ Alternatively. Then ó Î ò . Hence deduce that E 2 ¥ { ò õ E ò ô C õ E 7 2 " ô 2 2 2 ò ¥ õ f f ò ô ¥ ¥ 2 " " õ ò ò " ô f õ õ f " C · · ô { ô " C " ó Î ð f f d ò ò ò · õ õ 2 f ô ô f f E f { ¥ ò ó " Î õ ò ô · C { ¨ ¨ { ó Î ò must be positive. Thus each has the same derivative apart from a sign.
11: The graph of . We leave the others to be done as exercises. has an inverse function. { g " 2 f E " 2 f ó .11). The range of the restricted function is . at multiples of . C " ¥ ¥ ó ó " " ó . ð " ó ô d d M " " ò . To compute the derivative of we use Theorem 10. we deﬁned four other trigonometric functions— . 2 ó . E 2 C ð E ¥ ó 2 . The only one we consider here is the inverse tangent. the graph will have vertical asymptotes at the points where .2: 2 g ¥ u ó 2 . suitably restricted. The formal deﬁnition is the following. Each of these. so there is an inverse (denoted by ) with domain and range (Figure 10.5 OTHER INVERSE TRIGONOMETRIC FUNCTIONS In Chapter 7. Since the function is not 1–1. E 2 ¥ E E ó 2 2 . We denote are 2 E 2 ¥ ó . ' f U f . DEFINITION 10.OTHER INVERSE TRIGONOMETRIC FUNCTIONS 221 10. C C 2 . by 2 ¥ ó . g ~ u % 2 u 2 g ò ~ õ ô f g h ~ u 2 u g ~ 2 f ¥ ó . C ¥ ó . and . . C ó . we need to restrict the domain in order to construct an inverse function. Since . E 2 C 2 . The inverse function relations 2 ¥ ó ' . C g ¥ f ó . . õ ô 2 ô ò 2 ô ò g õ ~ ô 2 ~ ô M 2 ò ó Î ò 2 h ó .5 The inverse tangent The inverse of the function g u 2 u g f 7 2 ó . The natural choice is to take . u h E g g 2 C f 2 ¥ 7 7 is called the inverse tangent function. Figure 10. that is.
EXAMPLE 10. The point of this discussion is not to give practical methods for calculating function values.5. Then the form 2 2 ó F g ¥ ó F g 2 ¥ ó F 2 ó Î ò { 2 { ó Î ò We conclude One minor point: How do we calculate the values of and ? In principle.1 Computing inverse trigonometric functions ¥ Apart from a few special angles. . These functions are built into the calculating device precisely because they arise in a large number of applications. which does not have some mysterious method of its own for doing the calculation.15) and ¤ d 2 ¥ ó F 2 ¨ " ° (10. We put g and use Newton’s method in f ¤ 2 Ú ¤ . so are the values of the calculator. In . say . nobody would routinely calculate these functions in such a laborious fashion. To compute a value of we have to solve the equation for the value . but this is not true of the inverse trigonometric functions.5. 10. Put .2 CONCLUSIONS 2 ó F " 2 a % ¥ " ó " F g ¥ ó F 2 ô w ò The signiﬁcance of the inverse trigonometric functions really lies in their derivatives. so take as a starting value something a bit bigger than g f 2 ó F E E ¤ ¤ E 2 2 C 2 C ð C .16) These integrals arise in many problems. As we have already seen. since they are built into any reasonable calculating device.222 INVERSE FUNCTIONS 10. Compute Solution. such as Newton’s method to . the method described above is essentially that used by fact. These are ¤ d 2 ¥ ó Î ò 2 ¨ " ° (10. The derivative of a trigonometric function is another trigonometric function. the ﬁrst arises in oscillation problems when we use energy methods to ﬁnd the solution. compute values of the inverse trigonometric functions.11 to 4 decimal places. for instance. we have to use inﬁnite series. Our discussion in this chapter yields two results that we could not have obtained in other ways. However. we have to resort to numerical methods. " 2 U ~ r r ' " " " g a a ! g % % % " " " " " " " " g g g g § f f f f 2 2 2 2 § 2 2 2 2 ô ô ô ô w w w w ó ó ó ó ò ò ò ò F F F F 2 2 f f f f f d § " " 2 2 2 2 § ° 2 2 2 2 " E U ~ C ó F We know that Then . but rather to lay bare the processes which have to be taken into account when designing algorithms for calculators and computers. Of course.
¥ ¥ ¥ 19. 2 ¥ 2 2 28. E § ò õ ô C ó Î ò ¥ ¥ ÿ ÿ þ þ E % ó . E 2 C ¥ d 2 " · E d 2 C f r 2 ¨ Ê r É ó Î ò E 2 C ¥ d 2 f " · ó E 2 C ó . E 2 ò õ ô C ò õ ô E 2 ò õ ô C ó . . C ò õ ô E % ò õ ô C ó . ¥ 1. ¥ ¥ Differentiate the following functions: 23. 9. 223 . g ò õ ô ò õ ô ¥ ¥ ° ' ó . " E 2 ó Î ò C ó Î ò ¥ 22. 18. 6. 11. . 20. E 2 ò õ ô C ò õ ô E 2 ò õ ô C ó Î ò E 2 ó . calculate the following: 17. ¥ ¥ 26. C ó . 29. E 2 C ¥ E 2 ó . calculate the following: 14. 25. ¥ ¥ ¥ 2 If " u 2 u 16. 7. ¥ 30. ¥ If % w 13. E 2 ó . C ò õ ô E 2 ó . { 2 ò õ ô Ó E 2 C Ê 2 E " f 2 C ò õ ô E 2 C É ó . ¥ 24. E d 2 C d r 2 ¨ Ê r r É ó . ¥ ¥ % ò õ ô ¥ 4. Differentiate Hence evaluate the integral ° Î ò E C ¥ % 12. E 2 C ÿ ' þ 2 f U ÿ g þ ó Î ò ' d 2 f · { ó Î ò f { ¥ E 2 ó Î ò C ó . 8. 27. E ò õ ô C ó Î ò E ´ ó Î ò C ó Î ò ¥ " ó Î ò ò õ ô Determine the exact value of the following: EXERCISES 10. ¥ 5. Differentiate Hence evaluate the integral ° 31. C ó Î ò E " f C ó . C ó . 21. 2. ¥ 10.5 OTHER INVERSE TRIGONOMETRIC FUNCTIONS 3. C ó Î ò E " ó . 15.
The collection of functions we have had to introduce includes algebraic formulas. as we have seen. together with all combinations of them. are called elementary functions. If you think about the links between complex exponential functions and the trigonometric functions. but there are problems for which these functions are insufﬁcient to provide a solution. These are the hyperbolic functions and their inverses. Its auxiliary equation is with roots . there is another group of elementary functions which play a sufﬁciently important role in many problems to have been given speciﬁc names. the solution is (11. roots. trigonometric functions and inverse trigonometric functions. growth and decay and oscillation problems.1 HYPERBOLIC FUNCTIONS The hyperbolic functions are deﬁned as certain combinations of exponential functions. In real form. exponential and logarithmic functions. so the solution is Ê 2 g Ê g { . you will not be surprised to discover that the hyperbolic functions have many properties which are similar to trigonometric functions. not in the sense of easy.1) which we can write in complex form as as ¦ ) ¥ Ó f ¦ ) Ó É ¦ d ¦ ) ¥ Ó d ¦ ) Ó É ¥ 2 ® % % { " d d ð ð { 7 ) ó Î ò ¦ d ) ò õ ô ¥ { (11. The circular motion in the Tower of Terror provides such an example and we shall consider it in Chapter 13. As well as the elementary functions we have already considered. Indeed. .2) Now let’s keep this example in mind and consider the case % { f ð ð { ¦ " ¥ Ó ® ¤ d ¦ Ó £ { % " f The auxiliary equation is with roots . These functions provide solutions to problems such as projectile motion. 11. The word elementary is used here in its sense of fundamental or basic. which is an example of To motivate the deﬁnitions let us consider the equation undamped simple harmonic motion.CHAPTER 11 HYPERBOLIC FUNCTIONS Much of this book has been about the need to introduce functions in order to solve particular problems. there is nothing particularly easy about most of these functions. These functions.
We give names to the two combinations of exponential functions occurring in equation (11. There is a marked similarity between the form of this last result and equation (11. Ó Ó ¤ E ) d ¦ r ÿ ¦ d ³ Ù D Ó ó ¦ Î £ ³ ò C Ù ¦ % § ¦ ¥ ¥ Ó d Ó g d ) U r ¦ þ ³ d { Ù D Ó ò g ` õ ô d £ ¥ ¥ à Á The auxiliary equation is with roots . They are often pronounced as cosh and shine (or sinch by Americans). The similarity between the form of equation (11.4) and that of equation (11.1 We consider a case of strongly damped motion given by the equation % { U d ð { ` d ð ð { ) D ó Î ò ) ) We can express this in terms of hyperbolic functions by writing £ d ¦ ³ Ù ¥ Ó d ¦ ³ Ù Ó ¤ d £ ¦ § ¥ ¤ f ¦ where and . but this is simply because of the notation. However.2) by rewriting it as follows: ¦ ¥ Ó ¤ i Â ÿ ¦ ³ Ù 7 ¥ Ó á f ¦ ¦ ³ r ³ Ù Ù ¥ Ó Ó þ ® g E f ' ¦ ¤ ³ ¦ f ³ Ù g f Ù ¥ d ¤ É d ¦ Ó £ d £ É d Ó Ó Ó Ó d É g g £ £ ) É C ¥ { ¤ D ¤ ò õ d d ô { where and . so the solution is ¦ Ó Ê g ¤ f g ¤ É d ¦ ¥ Ó Ê Ê ¦ g £ ¥ Ó f g ¦ f ¥ g £ ¦ Ó Ó É g g É f d ¦ E Ó ¦ Ê ¤ ¦ ¥ Ó ¥ f Ó ) Ê Ê g £ D f C g g ¤ ó ¦ Î Ó ò d d É Ê { g g ¤ ¦ ¦ ¥ É Ó d g d d Ê ¦ ¦ ¦ Ó Ó £ ¥ Ê Ê É Ó C g ¦ ¦ g g E £ d § § ¤ ¥ ¥ ¦ (11. this notation has been chosen for a reason.3) becomes ) D ó Î ò ¦ d ) D ò õ ô ¥ is called the hyperbolic cosine of .1) is obvious.226 HYPERBOLIC FUNCTIONS We can regroup the exponential terms in a similar way to that of equation (11. It turns out that there are many similarities between the trigonometric functions and the hyperbolic functions. some of which we shall explore below.3) £ £ ¥ ¥ (11. while is called the hyperbolic sine of The function . EXAMPLE 11.2).4) .3) by deﬁning ¦ ¥ Ó d ¦ Ó ) D ò õ ô ¤ f ¦ Then the solution (11.
PROPERTIES OF THE HYPERBOLIC FUNCTIONS 227 In general. These are " 2 D F õ ô 2 D ó F Since the hyperbolic functions are combinations of exponential functions. then the solution can be expressed in the form (11.2 Solve the equation The auxiliary equation is By analogy with the trigonometric functions. we deﬁne four other hyperbolic functions. For the hyperbolic cosine. % { f ð ð { % { g f ð ð { 1. They are shown in Figure 11.1 Find general solutions of the differential equations in exercises 1–4 in terms of the hyperbolic sine and cosine functions.2 PROPERTIES OF THE HYPERBOLIC FUNCTIONS It follows from the deﬁnitions that and . 11. 2 5. If a heavy cable such as a power line is strung between two points. if the auxiliary equation of the differential equation % { ¤ d ð { d ð ò U ò g 7 7 õ D f 2 2 2 ô " ò ð E D D D ð õ " £ ò ò 2 ô { ó C ð { Î õ õ ¦ D ò ô ô g £ ò % r ( EXAMPLE 11. EXERCISES 11. Verify that the two sets of answers you have are equivalent. they are a very useful convenience in many circumstances and we shall use them as the need arises. Find general solutions of the differential equations in exercises 1–4 in terms of exponential functions. the curve it hangs in will be a hyperbolic cosine. we could still do all of the problems in this book without introducing them. of and The derivatives of these functions show a similarity to those of the trigonometric functions. 2 D " ó Î ò 2 2 sech csch i 2 D ó ) F g D ó Î ò ¤ g d ® . However. { 2. The general solution is therefore E ) ( D ó Î ò ¤ % d ) ( { D Ó E ) C { U f ® 1 has real roots .5) . The graphs can be plotted from a knowledge of the graphs . we have E % ¥ Ó d % Ó D C ó Î ò ¨ " õ D ô ò C õ ô ¨ ¥ Ó Ó % 2 g f ð 2 E g ¥ f Ó ð ð f 2 2 D Ó ¨ C ó Î " g g 2 ¨ % { r d ð { ` d ð ð 3. 4. it has roots ) { % .1. There is a simple occcurrence of the hyperbolic cosine that we see almost every day. These are the only values of these functions which can be computed exactly.
C ò " ò ó 2 õ Î 2 õ f ¨ ò ô ¨ ¨ ô { { g f ' f 3 . 3 If the trigonometric identity contains the product of two sines. unlike the corresponding trigonometric result. then since 7 ~ ~ ó Î ò ~ ó . . ó Î E ò { Figure 11. For instance if a term such as appears in a trigonometric identity. Be aware that the product of two sines may be disguised. We ﬁnd 7 2 . change the sign in the corresponding hyperbolic identity. ò õ ô There is a rule for going from a trigonometric identity to a hyperbolic identity. ò õ ô E 2 .228 HYPERBOLIC FUNCTIONS ¨ Notice that. ó Î ò f 2 . cos to cosh. U 2 ó f f f 2 2 " . etc. The result for the derivatives of the hyperbolic sine and hyperbolic tangent are obtained similarly. there is no minus sign. ó Î ò C ¨ The hyperbolic functions satisfy many identities similar to those of the trigonometric functions.1: Graphs of and P g " 2 f " " g ' g ' . We just use the deﬁnition: ä E ¥ Ó f Ó C â f ä E ¥ Ó d Ó C â 2 . EXAMPLE 11. õ ò 2 ô õ ô . . ó . ~ i " 2 sech 2 . 3 we will have to change the sign in the corresponding hyperbolic identity. ó Î ò { 2 ò . as long as differentiation and integration are not involved. Suppose we have a given trigonometric identity.3 Show that Solution. . The rule for getting the corresponding hyperbolic identity is this: Change all trigonometric functions to the corresponding hyperbolic functions—sin to sinh.
PROPERTIES OF THE HYPERBOLIC FUNCTIONS
229
EXAMPLE 11.4 We have so the hyperbolic identity is
i ¤ , ó Î ò £ , ó Î ò d ¤ , ò õ ô £ , ò õ ô E ¤ d £ C , ò õ ô 7 ¤ " 2 d 2 , 2 , ó ó Î , ò . ó Î f f ò ó " g Î 7 ò 2 7 % E 2 " " 2 , 2 2 7 , ó ó f Î Î g ò % " 7 C ò ò õ " 2 , , , E ô ò f ó ó 2 f õ Î Î { ò ò ô , £ ¤ ò õ ô £ ò õ ô E ¤ d £ C õ ô ó f ô f Î d ò { { { ò , , , õ ò ò ò 2 ô
This rule is a consequence of the fact that both the trigonometric functions and the hyperbolic functions can be expressed in exponential form. We have
¥ Î ò ó ó ò Î Î Ó f ò ò ¥ { { { g Ó f 2 f f , C d , , g f ó ó ó { Î 2 Ó
and
Ó
So, for example, since we have
, ó d 2 , ò õ 2 ò õ ô
Hence where . We have here another example of results about real functions which are explained by the use of complex numbers. Without complex numbers, the similarity between trigonometric and hyperbolic functions would have been inexplicable. EXERCISES 11.2 Prove the following identities. 1. 5. 7.
ò . ô 2 { , , ò õ ô
2.
2 , ò õ ô "
,
ò
õ
ô
2
,
ó
Î
ò
g
2
g
,
ó
Î
ò
10.
Î
ò
2
,
ò
õ
ô
õ
ô
2
,
ó
Î
ò
E
{
f
2
C
,
ó
Î
ò
9.
Î
ò
2
,
ò
õ
ô
õ
ô
2
,
ó
Î
ò
E
{
d
2
C
,
ó
Î
8.
ò
2
,
ó
Î
ò
õ
ô
2
,
,
ó
ò
õ
.
ô
f
E
{
E
E
2
f
2
f
f
2
C
C
C
,
,
,
ò
ò
ó
õ
3.
4.
6. sech
11.
ó
Î
,
ò
õ
ô
2
ò 2
%
,
ò õ
õ ô
ô
230
HYPERBOLIC FUNCTIONS
and for all
.
h
Find the derivative of the following functions.
§ 2 , ó Î ò E 2 C E " d 2 g C , ò õ ô E 2 C
13.
14.
27. A body of mass falls from rest under the action of gravity. It encounters an air resistance proportional to the square of its velocity. Satisfy yourself that the motion is described by the initial value problem
7 % E % C 0 7 0 ± f ) 0 ¨ ¨
is a solution of this initial value problem.
)
(b) What value does the velocity approach as occurring here?
E ) C º
gets large? What physical phenomenon is
11.3 INVERSE HYPERBOLIC FUNCTIONS
The hyperbolic functions have inverses which are useful for computing integrals. The function is not 1–1, but if we restrict the domain to , we do get a 1–1 function. The inverse of this and is called (somewhat incorrectly) the inverse hyperbolic restricted function is denoted by cosine function. It has domain and range (Figure 11.2).
2 , ò õ ô E l E 7 % l 7 % 2 ¥ E , ò l õ 7 ô "
)
(c) Let
E % C º
be the distance of the body from some ﬁxed point at time . If it is known that , ﬁnd an explicit expression for .
º º
X
)
±
Ý
(a) Show that
W , ó . ± Ý 0
±
where
is a constant which depends on the properties of the particular body.
g
,
ó
.
g
E
2
C

2
r
,
ó
Î
ò

ó
Í
E
2
C
25.
2
26.
"
f
2
,
2
ó
,
Î
ò
ò
õ
ô
E
2
C
E
2
ó
Î
ò
C
,
ó
Î
ò
E
2
C
23.
24.
Ê
"
2
É
,
ó
Î
ò
E
2
C
2
ö
õ
Í
E
2
C
21.
sech
22.
2
ö
õ
Í
,
ò
õ
ô
E
2
C
§
2
,
ò
õ
ô
E
2
C
19.
20.
2
,
ó
.
f
2
E
2
C
2
§
,
ò
õ
ô
E
2
C
17.
18.
2
g
,
ó
Î
ò
ó
Í
E
2
C

2
,
ó
.

ö
õ
Í
"
g
E
2
C
15.
16.
2
2
,
,
ó
ó
Î
Î
2
ò
ò
,
f
d
ó
Î
ò
2
2
f
,
,
ò
ò
2
õ
õ
,
ô
ô
ò
õ
ô
¤
¤
E
E
2
2
¥ , , Ó ó ó
Î
Î
ò
ò
f
d
2 2 2
, , , ò ò
ó
Î õ õ
ò ô ô C C
d
2
,
ò
õ
ô
g
Ó
12. Show that
and
. Hence show that
2
7
INVERSE HYPERBOLIC FUNCTIONS
231
We have the pair of results
¤ d 2 ¥ D ò õ ô 2 ¨ " °
Notice the difference between the integral provided by equation (11.6) and the integral that occurs in
"
f
"
E
2
f
E
"
¥
2
"
D
f
ò
"
E
¥
õ
"
2
ô
E
E
2
C
C
2
"
2
2
¥
C
D
D
ð
"
f
E
ò
ò
¥
C
"
"
õ
õ
D
ô
ô
C
¥
2
ó
ð
Î
Þ
Þ
ò
C
E
l
7
7
"
E
"
since . The domain of the inverse function is
¥ ¨ % w 2 D ó Î ò
, so for
7
"
f
2
D
%
ò
2
õ
w
D
ô
ó
Î
2
·
ò
E
2
2
2
D
C
f
ð
ò
"
õ
ô
2
D
ò
õ
ô
E
C
2
C
2
E
¨
2
To ﬁnd the derivative, let
for
. Then
we have
2
¥
D
ò
õ
ô
{
%
w
2
7
2
D
ò
õ
ô
{
Figure 11.2: Graphs of
and
P
U
2
'
¥
D
ò
g
õ
ô
"
{
¨ " " g ' U f " f 7 2 D ò õ P % ô w { 2 g "
¥
D
ò
õ
¨ ô " " g ' U C f " 2 ¨ ¨ f g f ' f
(11.6)
232
HYPERBOLIC FUNCTIONS
connection with the inverse sine (equation (10.15)). The integral
2 ¨ " °
Thus
E 2 C ð E ¥ C E 2 ¥ D ó Î ò C ¨
2
D
ó
Î
ò
E
2
C
2
D
ò
õ
ô
E
2
¥
D
ó
Î
ò
2
C
E
"
2
D
D
E
ó
ó
Î
Î
2
¥
ò
ò
2
"
2
C
¥
"
d
d
"
d
D
D
"
"
ò
C
ò
"
ð
õ
õ
Þ
·
ô
ô
E
2
C
ð
2
¨
2
¥
D
ó
Î
ò
The derivative of Then
is found in a similar manner to that of
2
¥
D
ó
Î
ò
Figure 11.3: The graph of
. Let
¥
D
ó
Î
ò
is used on the interval . The hyperbolic sine is easier to deal with because it is 1–1, so it has an inverse, denoted with domain and range (Figure 11.3).
E " 7 " f h C
¤
P
d
2
2
¥
'
D
¥
ó
ó
Î
Î
ò
ò
g
{
2
¨
"
2
f
"
E
"
¨
l
"
"
g
'
g
7
"
"
f
f
f
C
f
°
g
E
"
f
f
7
'
l
f
f
C
is used on the interval
or
, while the integral
"
f
2
,
U
f
.
INVERSE HYPERBOLIC FUNCTIONS
233
Thus we have the pair of results
¤ d 2 ¥ D ó Î ò 2 ¨ " ° 7 " d h " " f h " " ° ¨ ¨ " " " " g g E " 7 f f f " 7 f " 2 2 " f g f d C " f " " f ' E f 2 E 2 U ¥ f D ¥ g ó D f r ó F Î C ¥ ò f C 2 D ¨ ¨ ó 2 ¨ ¨ F
The derivative is found in an analogous way to those of the hyperbolic sine and cosine (Exercises 11.3). We have the pair of results
¤ d 2 ¥ D ó F 2 ¨
The inverse trigonometric functions and inverse hyperbolic functions provide a complete set of integrals for functions which are reciprocals of quadratics or reciprocals of square roots of quadratics.
2
¥
D
ó
F
Figure 11.5: The graph of
E
2
"
The hyperbolic tangent is also 1–1. It has domain and range inverse, denoted , with domain and range (Figure 11.5).
7 " f C
P
¥
D
ó
F
2
D
ó
{
"
F
Figure 11.4: The graph of
P
U
2
D
ó
'
F
2 g { 2
(Figure 11.4). It has an
E 2 D ó Î ò C ¥ ó F E 2 C E 2 ¥ D ó F C ó Í E 2 C 7. 8.234 HYPERBOLIC FUNCTIONS This is one reason for the importance of these functions. so that the two sides differ by a constant. 6. 13. it is not surprising that the inverse hyperbolic functions can be expressed as logarithms. 10. 11. Prove the following two results 2 and then solving the equation Ê " 2 2 d f d 2 " " É · ó Í ¥ D ò õ ÿ ô 2 E ¥ 2 2 D C ó f F " 7 þ ÿ " f 2 2 · f " d 2 þ ó Í " f 2 2 ¥ D · ò 2 õ ¨ ô ¥ D ó Î ò E 2 C 9. Show that the constant is zero. 2. E " d 2 C ¥ D ó Î ò E 2 C 2 g ¥ D ò õ ô E 2 C 1. Since the hyperbolic functions can be expressed as exponentials. The following list presents the results we have derived in slightly generalised form useful in applications. § E 2 ¥ D ó Î ò C E 2 C E 2 D ò õ ô r C ¥ D ó Î ò E 2 C 5. { { D ó Î ò 2 2 ¥ D ó Î ò { (b) Prove this result by letting terms of .3 Find the derivatives of the functions in Exercises 1–10. 2 ¥ D ó Î ò d 2 ¥ D ò õ ô E 2 C 2 ¥ " D ó F E 2 C 3. ¤ ¤ d d r 2 r 2 ¥ ¥ D D ó ò õ F ô " r 2 ¨ 2 ¨ r 2 f " " f 2 r ° ¤ ¤ ¤ d d d r 2 r 2 r 2 ¥ ¥ ¥ D ó ó ó Î Î F ò ò " r 2 2 ¨ ¨ 2 ¨ 2 2 2 f d " " " d r r r ° ° for in . ° ° Table 11. Prove the following two results stated in the text: ¤ d 2 ¥ D ó F 2 ¨ " ° 7 " E 2 ¥ D ó F C ¨ 12. For the hyperbolic sine we have " d 2 · d 2 ó Í 2 ¥ D ó Î ò (a) Prove this result by showing that the derivatives of the two sides of the equation are equal.1: Indeﬁnite integrals EXERCISES 11. 4.
1 as was done in Chapter 9. It is also built into software packages such as Mathematica. for example. roots. However. if we let o ¥ Ó E 2 C ð û ð û û 7 ) ¨ o ¦ ¥ Ó û ° E 2 C C û then There is no generally accepted name for the function deﬁned above. exponential and logarithmic functions. If is an elementary function. Try the Mathematica instruction 1 Another option is to deﬁne the function by a suitable power series. hyperbolic functions and inverse hyperbolic functions.CHAPTER 12 METHODS OF INTEGRATION 12. We deﬁne the generalised error function by ) ¨ o ¥ Ó g E 2 o û ¥ Ó E 2 C ð û Erf ° We can now write so that E 2 C Erf g The function Erf is one whose values can be calculated by numerical integration. For example. one option is to deﬁne a new function in terms of an integral. we might expect to ﬁnd another elementary function such that but this cannot always be done. say. This is in contrast to our discussion in Chapter 9.1 INTRODUCTION In this chapter. that there is no elementary function such that case. there is a name provided for a closely related function. 7 E 2 C Erf g ¦ E 2 C ° ) ¨ o E ¦ 2 ¥ C Ó û E 2 C . need not have a primitive function in terms of a given class of functions. where the main emphasis was on using numerical methods such as Simpson’s rule to evaluate deﬁnite integrals. Recall that the elementary functions are those functions which include algebraic formulas. we will look at methods of evaluating deﬁnite integrals by ﬁnding a primitive function of the integrand. trigonometric functions. . inverse trigonometric functions. In such a It is known. One matter that needs to be emphasized is that a given function .
is 2 r 2 2 ¨ E 2 C ß ° £ 2 2 ¨ 2 ¨ E 2 C E ð 2 C d ß " is a given positive function. If the graph of is rotated about the The length of a curve E E C 7 C E E r C 7 r C E 2 C { The length of the curve from the point ° to the point is Work û The work done by a constant force on a body moving in a straight line is the product of component of the force in the direction of movement. and . Although not classed as an elementary function. For these reasons. The volume of this solid is 2 ° r è % w E 2 C . it is conceptually no different to functions such as .236 METHODS OF INTEGRATION Plot[Erf[x]. A simple integral such as far more efﬁciently evaluated by making use of primitive functions. all deﬁnite integrals and functions deﬁned by deﬁnite integrals can all be calculated numerically and. and so on. the function ) ¨ ) ß À ) ¨ ) " d " ° E 2 C 2 ó Î ò 2 ö õ Í is just another way of writing so that is not a new function.5}] to plot its graph. but we will brieﬂy summarise them below. However. · ß . the axis and the vertical lines . their graphs plotted without making use of the concept of an indeﬁnite can be integral. Thus ¨ ¨ û © © r The area under a curve has already been dealt with in Chapter 9. the total distance moved. We will not consider such applications in detail in this book. For example. this is not always necessary or efﬁcient. Areas 2 ¥ ó F Volumes of revolution 2 for Let be a function such that then a solid is generated.5. if need be. volumes and so on. If area bounded by the graph. the . the for and in the domain of axis. in that there is a rule of calculation which produces an output number from a given input number. It is also helpful to know whether the functions deﬁned by an integral are new or whether they are known functions in another form. There are also practical problems such as the length of a curve. The worst that can be said about this function is that it is unfamiliar. we need methods for ﬁnding elementary functions for integrals when this is possible. areas.{x. In principle. which can be expressed as deﬁnite integrals.
then m. it can be shown that be a primitive function for . Now let E r C û f E C û THEOREM 12. Then .1 Evaluating deﬁnite integrals Let be a continuous function on the interval 7 r ° 2 ¨ E 2 C ß 2 r 2 where and are the end points of the motion. which is . for a falling body of mass . the force is a constant and in falling a distance the work done is . Then and must differ by a constant. For example. and hence .2 CALCULATION OF DEFINITE INTEGRALS There are two methods available to compute a deﬁnite integral. let be any primitive of . so we have ( ( 7 0 ( 0 12. û so that which is to say E r C û f E C û 2 ¨ E 2 C ° 7 E r C û f E C û E C E r C û f E C û ß ä E 2 C û â As a matter of notation. which we now state. if we can ﬁnd a primitive function for the integrand then we can express the deﬁnite integral in an exact form in terms of the primitive function. Thus . If we take the example of the Tower of Terror. related to the Fundamental Theorem of the Calculus.CALCULATION OF DEFINITE INTEGRALS 237 The work done represents the increase of kinetic energy of the body. However. and let 2 ± 7 2 ¨ E E 2 2 C C û varies. such as in a spring where û ° © û . Put . Then for all and are both primitives of . We have already considered numerical evaluations in Chapter 9 and in this chapter we want to concentrate on methods for ﬁnding primitives. There is a very useful result. Hence 7 ) ¨ E ð ) C û ß ° E 2 C û To prove this result. This represents the increase in kinetic energy. so that . we write for . If the force 0 U U 0 . We can do it numerically and (usually) approximately using a method such as Simpson’s rule. Then U " " ( ß ( g and hence m/sec. E 2 C E r C û f ¤ 7 E r C ¤ û d f E r E C 2 C û û % E 2 C ß r 2 7 û r g 2 ð so that ¤ d E 2 C û .
¤ ¤ ¤ ¤ d d d d r 2 r 2 r 2 r 2 ¥ ¥ ¥ D ¥ D ó ò ó ó õ Î Î F ò ò ô " r 2 2 2 ¨ ¨ ¨ 2 ¨ E r 2 2 % ò 7 2 f f d " " " ¤ õ " ô d r r d 2 f r 2 ò ò õ ° ° ° ° ô õ ô C f f g " 2 f ¨ % 2 k k ó Î ¤ ò ¤ 2 7 ¤ d 7 ¤ d ° EXAMPLE 12. we can rewrite the integral by using algebraic manipulation or trigonometric identities.1) ¤ 2 d 2 d ò d 2 õ ó  ¤ " ô ó 2 Î F Ú  d d ò f ¤ ö 2 õ Ó Í 2 2 ¨ 2 ¨ ¨ 2 2 2 2 2 2 ¨ ¨ ¨ ò ô ó ¤ õ Î w " 2 Ó ô ò ò 2 ° ° ° ° ° ° .1 Since (12. Firstly. to large books of tables. These can range from a small table. there are methods that make use of differentiation formulas and.1: Indeﬁnite integrals There are two main methods of transforming integrals into a standard form.1. There are two ways to do this: To make use of a table of integrals. such as Table 12. 3 Table 12. The calculation of a deﬁnite integral by this method requires us to be able to ﬁnd an indeﬁnite integral among the known functions. If you don’t understand why. try putting in the and see what happens. This chapter is largely about such methods.1). we may be able to rewrite it in terms of integrals that do appear in the table.238 METHODS OF INTEGRATION we have ä 2 ò õ ô f â 2 ¨ 2 ó Î ò ° Notice that it is unnecessary to put the constant in equation (12. 3 i ¤ ¤ If a particular indeﬁnite integral is not on our list. secondly.
INTEGRATION BY SUBSTITUTION 239 EXERCISES 12. as is done in the following example. EXAMPLE 12. 3. 2 ¨ E 2 ó 2 2 ¨ F ¨ 2 d 2 D " " ò d C õ § ° " j ° j é ° ° ° ° 2 ¨ 2 ¨ 2 E ¨ ò 2 õ U ô 2 2 d f " ó Î " ò 2 ' Ô j Ó ° Ò ° ° (12. 2 2 ¨ 10. then and the integral has the form ¤ d E E 2 C C û 2 ¨ E 2 C ð E E 2 C C ° Thus E E 2 C 2 ð C ð E E E E 2 C 2 C C ð C E E 2 C C 2 ¨ E 2 C Þ where is the Heaviside function deﬁned in Exercises 2. It is possible to use it as it stands. ° é ° 2 Æ ¨ ° § " 9. 2 ° . ô j ° 2 ° 12. we have û û ¨ This is the rule for integration by substitution . 8. 7. 12. 4.3 INTEGRATION BY SUBSTITUTION û This method is derived from the chain rule. is a primitive for .2) . Then for any differentiable 2 ¨ E 2 C Þ ¥ 2 ¨ Ó 11. g ° C ° ° ° 2.2 Consider 2 ¨ 2 ó Î ò 2 g ° 2 ¨ E 2 C ð E 2 C ó Î ò ° 2 g E 2 C ð 2 E 2 C If .6. Suppose function .2 Evaluate the following integrals: 2 ¨ E " f 2 ò õ ô g C 2 ¨ 2 ó Î ò 1. 6. 5.
3 In the integral 2 ¨ 2 ó Î ò 2 g ° û EXAMPLE 12. but then you should approach such problems as in the next example. there is no need to explicitly identify the functions . Then the i 2 ü Much easier! Note that after we have made the substitution the original variable . So putting all of this into equation (12.2) gives ° 2 ò õ ô f E 2 C û 2 ó Î ò ¤ ¤ d E d 2 C 2 ò ò õ õ ô ô f f 2 ¨ E 2 C ð E 2 C ó Î ò .240 METHODS OF INTEGRATION You may ﬁnd the above example confusing. so that by and by to get i E 2 C In this example. the function corresponding to in equation (12. EXAMPLE 12.4 In the integral 7 2 ¨ 2 d " · 2 ° 3 7 2 ¨ 2 ô w ò 2 ó F ° In the integral 3 ü ¨ 2 ¨ 2 ¤ ¤ d 2 d j ¨ § ÿ E 2 g j 2 § ü d § " ü þ C ¨ § ü ¨ ü ° 2 ü put ( integral becomes " d 2 ü also works). It is perhaps worthwhile to use the formula once or twice in this way to convince yourself that you understand it and why integration by substitution works. 2 . Thus and . In the integral we now replace 2 g ü 2 ¨ ¨ 2 ü put .2) is the sine function. so that and hence . and in order to use integration by substitution. In fact. the new integral must not contain ü ¨ 2 ¨ 2 g ü 2 ¤ ¤ ü d ¨ d ü ü 2 ò ò ó Î õ õ ò ô ô f f ° 2 ¨ 2 ó Î ò 2 g 2 ° ¨ 2 g ü ¨ Then symbolically write .
14. 10. so and the integral becomes 2 ° ° ° . 2 ¨ 2 2 d ó " F ° 2 ¨ E 2 ö " 17. 2 ° g ° ° 2. If tiable functions. 12. EXERCISES 12. 9.INTEGRATION BY PARTS 241 In these examples the substitution has been given. " ° ü ¨ ü D F õ ¨ 2 g 2 D g ó D Î ò ò õ ô 13. 16. 6. You will improve your ability to ﬁnd the correct substitution by doing the exercises—the more the better. õ Í C 2 ° 18. 3. 8. 5.3 Evaluate the following integrals: 2 ¨ 2 § ò õ ô 2 ó Î ò 2 ¨ E " d 2 C 2 g 1. ° and are two differen i º ¨ ¤ d º g d 2 ~ ¨ º ¤ 2 r U ¨ 2 d · d " ¨ E ¤ ü 2 § 2 ¨ d § 2 d 2 2 ~ ó d § · ü 2 ó º F ü Î r ò § § ° ° ° ° 2 ¨ 2 ô w ò 2 ó F 2 ¨ ° 2 ô w ò ü ¨ ) 2 2 ¨ ¨ 2 ó ¤ ¨ § ) 2 F 2 " ¨ ò ¥ o d d õ ¤ ô ) Ó r ü 2 put . ) ¨ § ¥ ) ) ö õ Í ° ü ¨ ü f ü 15. d " ° ¨ © Û Ó ê Ó E ó Î ò 2 ¨ 2 ö 2 õ Í 11.4 INTEGRATION BY PARTS Integration by parts is derived from the product rule for differentiation. f " 2 ò õ ô 20. 7. ô ° 2 C ° C ° 2 ° ° 12. 4. then E 2 C ð E 2 C d E 2 C E 2 C ð ç E 2 C E 2 C æ 2 ¨ ¨ 2 ¨ 2 ó Î ò f " · 1 ¨ g Ó " 19.
6 ° ° The same integral appears on both sides of the equation. The formula for integration by parts is easier to apply if we write it in the form 7 ü ¨ 0 ° f 0 ü 0 ¨ ü ° (12. So we are forced to take . We have to be guided in this choice by what we have to do with and .5 To ﬁnd E 2 C 0 2 ¨ E 2 2 2 2 C ð ó ó ó Î Î Î ò ò ò ß ß ß 0 Ó Ó Ó ¨ " " " 2 ¨ E 2 2 ¨ C ð 2 Ê ò õ ô ü ¨ ß Ó E 2 C ü where . and . EXAMPLE 12. we indeﬁnite integral for and .3) There are two integrals here.4) we put and . then we have to try another method. We combine the two integrals to get 7 2 ò õ ô ß Ó r d 2 ó Î ò ß Ó " 2 ¨ 2 ò õ ô ß Ó ° r d " É i Ê 2 ¨ 2 2 ¨ ò õ 2 ô ò ß õ Ó ô ß ° Ó r ° ¤ d 2 2 d 2 r ¨ ò 2 f 2 õ " g U ô 2 ó Î f ò ß ò 2 Ó õ ¨ 2 ô " ß ö 2 Ó õ f ß ö Í Ó É õ Í r 2 r r 2 " g g f f d ° EXAMPLE 12. With this choice of and . Since we need to integrate . it is no good taking because we have no .242 METHODS OF INTEGRATION If we integrate this equation we get 7 E 2 C E 2 C 2 ¨ ä E 2 C ð E 2 C d E 2 C E 2 C ð â ° which may be rearranged to give 2 ¨ E 2 C ð E 2 C ° f E 2 C E 2 C 2 ¨ E 2 C E 2 C ð ° (12. Application of the integration by parts formula gives have 0 ¨ ü 2 ¨ 2 ö õ Í 2 ¨ 2 0 " ¨ 2 ¨ ° 2 " f 0 ¨ 2 ö õ Í 2 " 2 0 ¨ ¨ 2 ö õ Í 2 2 ¨ 2 ° 2 2 0 0 ö ¨ õ Í 2 ¨ 2 E 2 ö ~ õ Í " C ü 0 ü ¨ ü It is sometimes necessary to integrate by parts twice. This is particularly common in integrals involving trigonometric functions. If we want to evaluate the one on the left and cannot do it. . then our hope is that we can do the one on the right. If neither integral can be found. .
2 2 2 ° 2 2 ° 2 2 ° 2 ° 12. equation (12. 3. In order to see where we are going. 11. 13. 9. 5.5) i 2 ò õ ô Ó d r d 2 ó Î ò Ó d r 2 ¨ 2 ò õ ô Ó 2 ó ¨ Î 2 ò ¨ ¨ C ¨ ¨ E 2 ö U 2 2 2 g õ Í ó ó d Î Î ¥ ¥ 2 ò 2 ò 2 ó ó ó Î Î ò Ó ò ° ° ° . 8.5 THE METHOD OF PARTIAL FRACTIONS The previous two methods of integration have been based on differentiation rules.THE METHOD OF PARTIAL FRACTIONS 243 so that ß r ß ß ° EXAMPLE 12. 14. 10. let’s ﬁrst consider an example. 12. 6. The method of partial fractions relies on using algebraic methods to rewrite the integrand in a form which is easier to integrate.7 ° ° EXERCISES 12. C ° F ° 2 ° ° 2.3) is often useful. i 2 2 ¨ ¨ 2 2 2 2 E ¨ ¨ 2 ¨ 2 2 2 ¨ 2 2 2 ¥ ¨ ¨ E ö ö ó 2 õ " Î ô 2 2 ¤ õ Í C ¥ ò Í w C ð ö 2 ò ó 2 d Î õ Í ò 2 2 ° ° ° ° f f 2 2 f ö ö E õ õ 2 Í Í C 2 2 2 2 ¨ 2 2 ¨ E ö õ 2 Í C ° 2 E ¨ 2 E C 2 The special case in equation (12.3) becomes (12. In this case.4 Evaluate the following integrals: 2 ¨ 2 ó Î ò 2 2 ¨ Ó 2 1. 4. 7.
8 above. This is called the partial fraction decomposition of the function. In the general case this may be difﬁcult and we will only deal with cases where the factorisation has already been done or is easy.244 METHODS OF INTEGRATION EXAMPLE 12. then the function can be split into separate fractions. each linear factor in the denominator will give a fraction £ d 2 in the partial fraction decomposition. ¤ d 2 d 2 r in the partial fraction decomposition. but we won’t worry about it just yet. In this case.8 It is easily checked that ' " d 2 f " " ¤ d 2 ' d 2 g U ¨ ° r ' 7 d ' d 2 2 We will have to consider methods of deriving the righthand side from the lefthand side. Let us ﬁrst consider rational functions where the degree of the numerator is less than the degree of the denominator. one for each factor. such as " Functions of this type are called rational functions . Each quadratic factor ¤ d 2 or 2 C E " d 2 C E g d 2 C E " gives i E " d 2 d ¤ d 2 E ¨ ' Ê d ' 2 C d ö õ 2 Í 7 f f 2 ¨ E " ¤ " g ' d d d f d d 2 2 2 2 2 2 d É C C U ö E 2 d õ " r d Í d £ £ 2 2 2 C d " U 2 d U d § d 2 2 2 7 ' d " 2 U d 2 d 2 . An example of such a decomposition appears in Example 12. 3 The second fact is that once we have factorised the denominator of a rational function. 3 We will only consider cases where the denominator of a rational function has factors which are all different. where and are constants. In each fraction the degree of the numerator will be less than the degree of the denominator. If we want to ﬁnd the integral g we can do it easily by writing " " ° g ° Partial fractions are used for functions which are quotients of two polynomials. There are two facts about polynomials which we will make use of: Any polynomial can be factorised into a product of linear and quadratic factors. where a fraction £ 2 r is a constant.
but in this example we will give more detail. Hence and .1 of Chapter 5). so we i g d E 2 " d " 2 C d ¤ 2 d E g E d g 2 C d £ 2 C E " 2 d ¤ so all we have to do is ﬁnd and the left hand side. With practice. so we can substitute any two values of to get two equations in and . To ﬁnd and we equate coefﬁcients of like powers of on both sides of the equation (Theorem 5. and . The above equation can be written so we have 7 % ¤ d £ 2 7 ¥ f £ d 2 E ¥ d ¤ C d 2 E ¤ ¤ d £ C U g " 2 g ¤ We could solve these equations simultaneously for quickly ﬁnd and .9 Find the partial fraction decomposition of 2 Solution. so £ " f " f 2 2 g g ¤ d " " f f £ 2 ¤ f 2 g f 2 g f 2 ¤ £ We multiply by the denominator of the left hand side to get E " f 2 C E ¥ d 2 ¤ C d E " d 2 C £ U Putting gives . In this case put to get and then put to get . We have 2 E " d 2 C U E " f 2 C 2 ¨ E " d £ 2 C ¥ ¥ U E " f d ¥ f £ ¤ 2 C ° EXAMPLE 12. you should be able to do this without expanding the right hand side completely. One way to do this is to multiply through by the denominator of g d 2 E " g d d 2 2 C E " E g d d 2 C 2 C E " d 2 C 2 C £ £ .THE METHOD OF PARTIAL FRACTIONS 245 EXAMPLE 12. Some choices of involve less work than others.10 Find . This gives . but we already know £ ¥ " d d 2 ¥ 2 ¤ d 7 " % U f £ Solution. The form of the partial fraction decomposition is 7 ¤ d £ 2 This has to hold for all values of . g f ¥ g f ¤ .
Solution. we have used the result in example 12. On dividing the numerator by the denominator.9. in the last step. The method here is to do a long division and then ﬁnd the partial fraction decomposition of the remainder. 2 ¨ E " § E " d 2 d " 2 d g d C E 2 2 C " g § E d " 2 C d 2 C ° or i where.246 METHODS OF INTEGRATION Now we can return to the integral: g d 2 g g ° U ° We have not yet considered the case where the numerator of a rational function has degree greater than or equal to the degree of the denominator. but we will show you how Mathematica can be used to ﬁnd the partial fraction decompositions. We illustrate this in the next example. EXAMPLE 12. integrals of the type 2 ° We won’t go into the details of how to attack these cases by hand. we ﬁnd 2 g d 2 ! d § 2 a d § 2 g The method of partial fractions can be further developed to deal with cases where higher powers of the linear or quadratic factors appear in the denominator. i ¤ d 2 ¥ 2 ¨ ó 7 E g F Ê g g " g d g d f 2 d g d 2 E C 2 2 2 E " d 2 ' " ¨ " d f d d Ê " d " 2 " 2 2 C C ¨ 2 2 2 d d ö § g õ f E d d Í " 2 2 " " " f d d d d f f E " 2 2 2 2 " " g g g g C f g E f f " 2 2 2 C ö d É É õ Í 2 C ° g g d 2 2 ¨ ' E d " d 2 2 C E " f 2 C . The relevant Mathematica instruction is Apart. so to ﬁnd the partial fraction decomposition of 2 we use Apart[x/((x+1)(2x+1)ˆ3)].11 Expand g d g 2 ! d d 2 ' § 2 d a d 2 § 2 g in partial fractions. for example.
d 2 C E " d 2 C 2 ¨ g f 2 ó 2 9. 6. ) ¨ " " f d § § j j ) 15.INTEGRALS WITH A QUADRATIC DENOMINATOR 247 Mathematica returns the answer g f g g d 2 2 " ¨ d 2 ¨ E ¨ 2 E " ü E ò 2 r ¨ " õ ¨ d E E 2 ô f ³ ` d 2 ó ' j 2 2 Î g 2 2 2 ò C C f g ¨ ¨ a d d " E E ü d d " " d " 2 g " 2 " 2 d C f f f d " " ò 2 2 f C C ° õ 2 2 d C C " 2 ô 2 2 2 ü ¨ § E ¤ ° ° ° ° ° ° ° ° 2 g d " f " ¤ d 2 d " C d 2 £ 2 2 r d ° " § ± ± § 2 ~ E ¨ " ¨ E g d ' d 2 2 d g ~ 2 2 EXERCISES 12. d ° 2 ° 2 ¨ 2 Î ò d " 7. ° 2 C ~ E g 2 11. 4. we often end up with integrals of the form These can be evaluated by ﬁrst writing the numerator in the form numerator ± (where is a constant) and then writing numerator denominator derivative of denominator denominator other terms denominator It should now be possible to perform the integration. The procedure is best illustrated by an example. 2 g ¨ C E E " g E 2 d f d 2 2 C C g E C r " d 2 g j C ° ° ° . ) 16.5 Evaluate the following integrals: ¨ " 1. ò õ ô 12. derivative of denominator other terms 2 ¨ ó Î ò g d ~ 13. d § 2 ° 8. Evaluate the following integrals by ﬁrst making a substitution and then using partial fractions. 14. § 2 ° 2 ¨ 2 g d " d 2 ' 2 5. 3. E 2 d " C ° 2.6 INTEGRALS WITH A QUADRATIC DENOMINATOR When applying the method of partial fractions. 10. 12.
a. there are many other types of integral for which a primitive function can be found. For indeﬁnite integrals we state the function to be integrated and the variable of integration: Integrate[f[x].13 Evaluate the integrals i " ¤ ¤ d d d 2 ü Ê Ê ¨ " g ü U g d ü The ﬁrst integral is a logarithm. Then we have .b}] ¦ ° ) ¨ o ¥ Ó and ³ EXAMPLE 12. we also have to state the range of integration: Integrate[f[x].b}] The instruction NIntegrate is used for deﬁnite integrals only. We use Integrate as follows. while the second may be done by letting " ° g d 2 g ° ' " d 2 ' ° 2 2 ¨ ¨ U r d d 2 E " " g " d ¥ d 2 d ó É 2 ü C F 2 " g ¥ ° ° ó É g g g g F f f f f f 2 2 2 2 E E ¨ ¨ ¨ ¨ r r r r r g d d 2 r d ¨ f d d g g 2 2 d 2 2 E g g 2 2 g d d g 2 g g d d ô g 2 2 d w d d d g g ò d 2 2 2 C C 2 g 2 2 C ö ö 2 õ õ Í Í ° ° g g g ' ' g g ' ' 2 2 ¨ ¨ r r d d " 2 2 g d g 2 d d ' 2 2 § .{x. The ﬁrst instruction expects Mathematica to produce an exact result. such as 7 2 ¨ " d 2 · ° or those involving powers of trigonometric functions such as § ° Usually Mathematica can handle such integrals without any problems.12 ° ° 2 ¨ 2 § ô w ò ° As well as the integrals we have dealt with here. namely Integrate and NIntegrate. There are two Mathematica instructions for integration.248 METHODS OF INTEGRATION EXAMPLE 12.a. Important classes are integrals involving square roots of quadratics. so we use the form NIntegrate[f[x]. usually using an indeﬁnite integral. If it cannot do this then the second instruction will evaluate a deﬁnite integral numerically.x] For deﬁnite integrals.{x.
If we can compute them exactly by using indeﬁnite integrals. Generally speaking. i a g g ` ) ! 7 ) ! d ¨ % " o ¦ ¥ Ó ) ¨ ³ o ¦ ° To evaluate 3 ÿ ÿ g 2 þ ó Î ò d ÿ g 2 þ ò õ ô þ ö õ Í d ÿ ÿ g 2 þ ó Î ò f ¥ Ø ÿ E Ó g 2 2 ³ C þ ó ò õ F ô E þ 2 C ö ô õ Í w ò f × d 2 ¨ " g 2 § ô w ò To evaluate 3 use Integrate[(Sec[x])ˆ3. Otherwise we have to use numerical methods to calculate them approximately. ° The result is 12.7 CONCLUDING REMARKS There are two main reasons for calculating deﬁnite integrals in terms of indeﬁnite integrals. is a new function.CONCLUDING REMARKS ° 249 The result given by Mathematica is.0.5}]. On the other hand.x]. then we do so. work) can be expressed as deﬁnite integrals. it is easier to evaluate deﬁnite integrals exactly if we can ﬁnd a primitive of the integrand. after some rearragement of the terms. the function ) ¨ o ¦ ¥ Ó ° E 2 C û 3 cannot be calculated in terms of the functions we have met earlier in this book. 3 Many problems require us to deﬁne functions in terms of deﬁnite integrals and it is often useful to know whether this deﬁnes a new function or is just another way of expressing a known function. the function ) ¨ " ° E 2 C û 2 ¥ ó F E 2 C is not a new function. volume. 2 ¨ 2 § ô w ò ° use NIntegrate[Eˆ(tˆ2). This is a fact which can be proved and it is not just that we are not clever enough to think of a way to do it. Many quantities (area.{t. For example. . Consequently. We know that .
ò 2 ô M " d . METHODS OF INTEGRATION 2. 2 10. ' " d 2 ' d 9. Ó " d 2 g 5. 2 8. 4. § U d 2 ó .7 250 1. ' d 2 g 2 ò õ ô 7. 6. U d 2 ô M ò 2 d 2 d 2 2 ¨ 2 ¨ ° ° " d 2 d 2 r d 2 U d 2 2 ¨ 2 ¨ ° ° r d 2 ó Î ò g f 2 ó Î ò r d 2 ` d 2 2 ¨ 2 ¨ ° ° " d Ó d Ó a f 2 g " d 2 U 2 ¨ 2 ¨ ° ° " d 2 d 2 % g d 2 ! d 2 2 ¨ 2 ¨ ° ° Evaluate the following integrals: EXERCISES 12. 2 " 3.
Since it consists of a mass constrained to move on a circular path inﬂuenced only by gravity. hyperbolic and trigonometric functions. together with their inverses and all possible compositions of them . We used these methods to deﬁne logarithmic. Figure 13. the equation for the Tower of Terror will apply equally to the simple pendulum.) A simple pendulum consists of a mass suspended from a ﬁxed point by a light inelastic string (Figure 13. exponential.CHAPTER 13 A NONLINEAR DIFFERENTIAL EQUATION In this chapter.1: The simple pendulum We have seen two methods for constructing solutions of differential equations. where is the radius of the circular part of the track. These functions. we will discuss some advanced topics which arise by considering the motion around the curved part of the track in the Tower of Terror. In Chapter 9 we derived the differential equation ~ ó Î ò ~ ~ ó d Î º ð ò ð ~ í ~ ò õ ô to describe the motion around the curved part of the track in the Tower of Terror. We will ﬁnd that the differential equation which governs this motion is also applicable to the case of the simple pendulum. namely power series and integrals. but the solution will have more physical meaning if we leave it in place. which we will also take as an example in our discussion.1). (We can remove the mass from the equation as we did in Chapter 9.
The differential equation we are now considering is such an example.1) term. the dependent variable and its derivatives occur to the ﬁrst power only. for example. The equation (13.2) As yet we have not imposed any initial conditions. It is called a nonlinear differential equation and the presence of is not linear because of the the makes it impossible to obtain a power series solution in the way we have done for ealier equations. This is simply a fact of life which we have to accept. As in Section 10. Such problems are inherently difﬁcult because their solutions are functions which are more complicated than elementary functions. There are. however. As before. Recall that 0 . in the case of forced damped oscillations.252 A NONLINEAR DIFFERENTIAL EQUATION are called the elementary functions . differential equations which do not have elementary functions as their solutions. and by the chain rule we have ~ ¨ ¨ ó Î ò º ð ~ Hence the differential equation becomes and so (13. but the solution is in terms of new nonelementary functions. we begin by multiplying the equation by to get ó Î ò ~ ó Î ò ~ and then multiply by to get Now we express the left hand side as a time derivative.3.1 THE ENERGY EQUATION We recall from Chapter 8 that a differential equation is linear if it has the form E ) C û ~ E ~ ò õ ô 7 % ð f 7 ~ C % ) ¤ Ê % ¨ ¨ Ê ð % ~ ~ ~ ò ~ ) ~ ~ õ ó ò ¨ ¨ ) ~ Î ó E ô ~ õ Î ¨ ò ¨ ) ô ò C ó ð E " g Î ð ¤ ~ ~ ò ~ É ò d ) º õ ) ð ¨ ¨ ¨ ô f ¨ ~ E f ) d d ð d C ð ð ~ ð ð E ð ð ð ð ð ) ~ ~ ~ C ~ ð ~ f ð d ð 0 ~ C ~ ð ~ ð ~ " g ¨ ~ E ¨ " g É ) C ) ~ r In a linear differential equation. With the elementary functions at our disposal. as we did. we can solve a large number of differential equations. Let’s suppose the body starts at the lowest point and so with velocity . The equation can be solved by integration. 13.
We assume that it is constrained to the track and cannot drop off the top portion. If reach the top—it cannot have negative KE—and so it will fall backward and oscillate. This energy would be released if the the initial level. If . ~ Ê Ê ~ C ~ É Ê ó C Î ò ~  ò B + õ ð ô B C B C C É  ¤  + The initial conditions then become equation (13. Using these initial conditions in (13.3) we obtain C ð ~ The next step depends on the values of the constants and in particular on the relation between and . Then For the Tower of Terror we have H C ð % H 7 C < C C K C M so that we deﬁnitely have . km/hr or 44 m/s and shall assume m. then it can be zero and this will allow to be zero. the body’s velocity is given by and so is its kinetic energy at time . which occurs at the lowest point. The second term on the left hand side is the potential energy (PE) and is the gravitational energy gained by the body rising above .4) . As it travels up the arc. At the top of the circle. then it is impossible for the body to KE when it reaches the top and so it will rotate.2) gives and ~  ~ + ð C ~ + + ð ~ . then the body will have positive . Mathematically this corresponds to whether the term C C M ~ C É K ó Î ò C C B C C can be zero or not. The expression is the initial kinetic energy (KE) of the body. The solutions are different in each case. If .THE ENERGY EQUATION 253 Hence or % ~ ò õ ô B + d ð ~ % B C ~ ò õ ô B ð ~ C where C This result is an equation expressing conservation of energy. If . The reason for this can be seen physically if we imagine the body to be going around a circular track. the height is and the PE is . and the times at which this occurs will be when the motion reverses in the oscillating mode.3) (13. Thus the above equation says that d + ¦ d KE + PE time ú ú KE PE initially ú ú From the energy equation (13. the body will just reach the top and then stop—a very unlikely outcome in practice. It is zero at the lowest point where body fell back to the original level.
5}.0. but they easily could if there were sufﬁcient demand. If we want to emphasize the value of .1.5) 7 ç E g ~ ~ C E ó Î g ò ~ C ± f ó Î " ò æ ± f " ð ~ ~ ~ g ~ ~ g ± Let and g w 13.5) in terms of by writing ± ± ± û ¨ û ± 7 E 7 t 2 C û E ± t 2 C û and so we get ¤ d ) g E C û A symbolic manipulation package such as Mathematica has many functions built into it besides the elementary functions.1 The case Then equation (13. where . Consequently and hence . Among these are the elliptic functions. using calculate the function . We can express equation (13. Here is a plot of the Mathematica instruction Plot[EllipticF[x. Mathematica can be used to . we need a function " û whose derivative is given by g ó 7 Î Î ò ò g ) ¨ ¨ ± ± f E f E " " ó Î C · ò · C " ± ) ¨ E f 2 " C ð g · ~ For convenience. we substitute .254 A NONLINEAR DIFFERENTIAL EQUATION and so ) ~ ¨ ¨ ~ " Þ (13.1]. the notation used is EllipticF[x. there is not just one function. Very few calculators have builtin elliptic functions.7) û There is no elementary function with this property and so we have to deﬁne the new function letting { ¨ " ° E 2 C û { 2 ó û In order to integrate the left hand side.5.1.{x. This function is called an elliptic integral of the ﬁrst kind . so that (13. In fact. Then the last equation becomes (13.PlotStyle>RGBColor[1. but rather there is a different function for each value of .0]] E " t 2 C û û % ¤ % E E % C C û ) g E C % û E % C % E % C The initial condition gives .4) can be written as by .6) The values of this function depend on the constant . we write or .b].1.
1 in two ways.1}] and gives the answer 1. If the graph of straight line (Figure 13. E ± t 2 C û ± E " t 2 C û Figure 13. We ﬁrst use the NIntegrate instruction of Mathematica and equation (13.{t.5 1 2 3 0.5 1 1.5 Note the similarity of the deﬁnition of to that of the inverse sine function in Chapter 10 and the is plotted for . Next.3). for small .5 EXAMPLE 13.5 1 0. Use Mathematica to evaluate Solution. In fact. ± " E ± " t { C ± û û 1. The full Mathematica instruction is NIntegrate[1/Sqrt[10.5 1 1.2: Graph of { ó Î ò f " ± · f " · . the function " E { C ð û has derivative 7 % w i E r % t 2 C { û Figure 13.5 Sin[t]ˆ2].THE ENERGY EQUATION 255 3 2 1 1.5] to get the same answer. we use Mathematica ’s builtin function EllipticF[1. for small .5 0. û E r % t " C û E r % 7 " C û By the fundamental theorem of the calculus. then the graph is almost a similarity of their graphs.3: Graph of 2 ó { Î ¨ ò ± ° Hence.5 1.0.6) to evaluate .5 1 0.5 0.5 1 1.0.5 1 0.08322.
we can compute function values of am from those of by using. In fact.4: The Jacobi Amplitude function.5 1 0.5 1 0.5 (page 253). Solution.4. using equa .8) 1 1. the Newton–Raphson process.2 Find the position of the car in the Tower of Terror after 0. We have 7 a ` " n 0 0 Ý g Ý and so U % ÿ U 7 E ) ! ! % C am g s g ~ To ﬁnd the time taken to traverse the curved portion of the path. we have Ê ) g É am and so Ê ) g É g g ~ am In principle. we need when ) ~ ö M î ' g g U % n Ê r % t þ r û g ! % ! " g § % g a ` ) " É am g radians r % ± r g n E where m/s and tion (13.1. Find also the time taken to traverse the total curved portion of the path.1].5 Figure 13. It is another nonelementary function.5.5}].7). we have ~ U U n 0 m.25 s. the function am is built into Mathematica and we can plot its graph as in Figure 13. The Mathematica expression is JacobiAmplitude. using the Mathematica instruction Plot[JacobiAmplitude[x. û 1 0. EXAMPLE 13.{x. Then . Using am in equation (13.5 0.5 1. for example. Hence. This means has an inverse. not familiar from elementary calculus.256 û A NONLINEAR DIFFERENTIAL EQUATION so is an increasing function and hence 1–1. We also found that û Ê ) g É ¥ û r g % C (13. This inverse function is called the Jacobi amplitude and is denoted by am.8).1.
the system oscillates and the solution is given by 7 E ) sn ± . We require a rigid rod rather than a string.1. and it is worth devoting some effort to understanding it. and sn is a function called the Jacobi elliptic sine function. you will ﬁnd that a full analysis of the simple pendulum is anything but simple.2 The case g u (13. . we can use equation (13. so that the pendulum is able to oscillate through any angle.9) i { ó Î { ò f ± { " f · " · £ . or even rotate completely around its point of suspension. at and it might oscillate. In fact. There is a systematic approach here which is applicable to more problems than just these three. but the functions which arise are different. Its where deﬁnition is in terms of the Jacobi amplitude am and is given by E E ± t 2 C C ó Î ò E ± t 2 C ~ E g C ~ ± sn am It is contained in Mathematica . In the case of low energy ( ). For various values of . where it is denoted by JacobiSN.9) to get graphs of the solution as a function of time. In spite of the name. A simple pendulum consists of a mass of negligible dimensions suspended from a ﬁxed point by an inelastic rigid rod of negligible mass. In this problem it is 7 { ¨ ° E 2 C û whereas in the air pressure problem it was { ¨ " ° E 2 C û and in the oscillation problem it was { ¨ ° E 2 C û The cases and do not arise in the Tower of Terror. let’s make a few remarks about the simple pendulum. There are also amusement park rides where a train oscillates or rotates on the interior of a circular track. The solution is then obtained by using the inverse of this function. and hence energy. namely that at the lowest ~ ± g C ¥ ó Î ò g E ) C ~ g g u g g w 13.THE ENERGY EQUATION 257 It is important to appreciate the similarity between the method of solution of this problem and the way we solved the oscillation problem on page 215 and the air pressure problem on page 210. The exact nature of the solution will depend on the initial conditions. the name is only there to distinguish it from the compound pendulum which we shall not consider. Its values can be calculated and its graph plotted as for any other function. we use similar methods to To solve the differential equation of motion for the case the previous case . The motion of the pendulum depends on its velocity. rotate or in the transitional case. Before continuing. In both cases the differential equation leads us to deﬁne a function with a given derivative by means of an integral. simply move to the top of the circle and remain there. but they do arise in other situations—for example in the simple pendulum. The difference is in the deﬁnition of the new function. We will simply quote the results and use Mathematica to plot the relevant functions.
4 0. the solid line is the solution (13. .9). Before plotting these graphs.10) ~ ó Î ò d ð ð ~ ~ ~ ~ ~ ó Î ~ 5 10 . and in the next two graphs. This is now a linear equation which we have seen (13.3 m/s.5 1 1. The approximation is excellent for the ﬁrst case. We expect that the amplitude of the oscillations will increase as increases. while the dotted line is the approximate solution (13. ó Î ò ~  + ~  + ð  + with initial conditions before.4 1. The equation of motion is % ~ ó Î ò d ð ð ~ which we write as If is small.258 A NONLINEAR DIFFERENTIAL EQUATION point.5 1 0. 25 m/s and 30.6: The case m/s. 7 H C ± % Here.5 m/s.10). and < = ± Figure 13. and this is indeed what happens. then ò and the equation is approximated by % ~ d ð ð ~ n How good is the approximation (13. but gets worse as we increase the initial velocity. .10)? We have drawn three graphs below for initial velocities of 8. the body has velocity . In each graph. we have used < Ý . it is interesting to look at a common approximation that is normally made for the simple pendulum. Its solution is and .2 0.5 5 10 < 7 < ± H C Figure 13.2 10 5 0.5: The case m/s. 0.5 10 5 0.
we mention the case when .7: The case m/s. In this case. but the principles used for deﬁning them by integrals and inverse functions are exactly the same as those for trigonometric and inverse trigonometric functions that arose from the problem on the Anzac Bridge. ó . indicating that the body takes forever to reach the top. Notice that asymptotes. . Here the body has just enough energy to reach the top of the circle. The graph is plotted in Figure 13. C ¥ ó Î ò g E ) C ~ 4 3 2 1 8 6 4 2 1 2 3 4 2 4 6 8 13. are horizontal . the functions which arise as solutions also become more complicated. ~ · and ~ ' ` % ~ r % ± r % ' 0 Figure 13. but not enough to go past it and rotate.CONCLUSION 259 2 1 10 5 1 2 5 10 For completeness.8.8: The critical case f where . The elementary function solution we obtain is 7 E ) . g Figure 13. Fortunately. much of the drudgery is now taken out of the calculations by tools such as Mathematica. and so is the function û " ± { ó Î ò { g ¨ f " · ° E 2 C û This function is an elementary function and can be calculated using suitable methods of integration. As practical problems become more complicated. Nobody has to do heroic calculations any more— computers are far quicker and more accurate—but we do need to understand the principles behind these calculations.2 CONCLUSION The functions we have used in dealing with this problem in circular motion may be unfamiliar. This is simply a fact of life.
( ) ) ) % ) ( ) ( % ( Answers 2. g d g ' ~ " 9. A reasonable domain for would be maximum sea depth. The points . so the domain of . There is only one value of for each value of . The values and are equal at all points . 20209 N/m Answers 1.ANSWERS Answers 1. (b) 88. are not in g ' % " ¦ ¤ 13.2 m ( m) % u g 2 ~ " " " g u w ~ " g g 2 2 2 f f u 7 a u % 2 2 2 f 7 § a E 7 7 " 7 7 a f r " " § d f d d ¥ d § 2 E C 2 2 2 2 2 g U g C U ~ © g U ¬ f f © © ¬ E " E E 2 d 2 ~ f " 2 C C C 12. " 2 2 " ® 2 ( k ( k 16. . 5. so h g 2 E 2 C E ( 2 C U g E a % " C © 15. 2. 3. 10.3 1. or g d 2 7 2 ~ " d " 7 2 f " 11. A reasonable domain for is all for which 3. (a) 45 m/s. . 19620 N/m . 7. 2.1 8. 4. A prediction for the population in 2000 is 20 million. 10200 kg/m 3. 6. but they are in the domain of and . . Automobile braking systems 2. The maximum sea depth is about 12 km. but it is unlikely that any normal submarine could dive to such a depth. means that there were 24 million people in the city in 1983. .6 1. Either variable could be taken as the independent variable. but (usually) two values of for each value of . U a ` g ! % r r ' ' U ~ r E ) f ' C ~ " d 2 g d 2 g 2 " d 7 d " 2 7 7 % 2 E 7 " 2 2 d d ~ " ¬ " § C 2 d 2 7 ' 7 " 2 ' d g f d g 7 ` 1.
. 20. 21. Þ . is the population and is some constant. ! 2 u r U ~ ' r E r C E U C ~ E C 7. 6.262 ANSWERS Answers 2.3 6. " d 2 g f 2 % " ¬ " ' ¨ 17. û 7 û (d) where is the force and is the universal gravitational constant. E l 7 ' f t h E l 7 % t " f ) and " 4. (or 11. % " 2 2 u " g u 2 u " E " w 2 l u 2 u " 10. 9. 8. ) % " 2 2 u " f % 7 7 2 2 f " f " f E 2 C 8. 17. h t g f 7 g k 2 E l 7 ' f t h 1. 2 . 12. (or 15. E l 7 g " g 2 E l 7 l f C g 2 E l C g 2 16. or E l 7 ' t E l 7 % E " 7 % t E " 7 % h t ¡ ' 7 ¥ " f 4. 2. 3. " " 2 2 7 7 " 2 f f " 2 © E ¥ 2 C 18. This is easier to see if you take the scale on the ' 2 % % " 2 % g 2 % 2. ± 9 7 ± 9 (c) where is the surface area. No such exists. 5. As one example r ' " % h E t g C h 7 ! ` ' % E " C 7 a % v ` % g E  r 2  % C t 7 % " k 2 5.4 3. is the radius and is some constant. ± ± © 0 p ï 7 0 7 ± © ± p ï 19. Use the ranges U " . 18. axis as multiples of 2 E l 7 g g f 7 l f C g 2 E " 7 a f C g 2 19. h Answers 2. 2. E ' 7 g g 2 E h g 2 l u 2 u l f ` u 2 g 13. is the velocity and is some constant. where is the kinetic energy. 2 Answers 2. (a) (b) where is the number of vehicles. All E % C All 6. 14. .2 E l 7 % t 7 l f C ¡ % h t ¡ " h E l 7 % t h 1. 7.
1 1. Y 20. converges to 4. 0 B + 21. Answers 3. Convergent. H 7 % 7 B % C % C B H % 7 C % H % < % % 7 % 7. 3 19. 13. 12.ANSWERS 263 Answers 3. Y 25. Conv.  + 24. d + 6.2 1. Div 22. The function is discontinuous at . 23. Conv. Divergent 15. ¤ B 3. H % 7 % 26.  d + Ú ¤ 2. 0 18. 11. H 7 C B C B B 7 B H B + < B . converges to 3. not possible + Y 14. d  B + 5. Conv. C 8. 9. 2 C + Y 17. H % 7 % % C 7 % 7 % C % C Á H 7 C % 7 = < C C % H H C % C 10. 4. Conv. Cost H C ¡ ~ B C Y ¡ ~ d C Y 27. Conv. limit 1 16.
% 4. The graph of U d 2 ' ' E E 2 C 2 ð C ð 2 U E 2 C 2 ð C ð E g d 2 C ~ " f E 2 C ð 7. 8. E 2 C ð 3. Both are differentiable at (and at all other points). 5. ( % 2 14. 9. ' g " " " g f E ) C á ³ R g © " E ) C Þ 28. " d % " f ! g " d % " d ! g 31. " " g " " f f g g f f is shown in the . 12. 11. E g ) f C ¥ " Þ 29. The function diagram: is differentiable at all points except .264 ANSWERS Answers 4. 32. The diagram shows . 2. 34. " E 6.1 ~ " f ` ' r g ~ " f ' g ~ " ! g U 1. ¤ 2 § " f ' d ' f " 33. The graphs of and are shown in the diagram. ð % 2 2 " { " " g f f " f g f ð 13. 10.
length of the river U g 7 U U " 7 a 13. (b) The quotient represents the approximate change in the number of people per unit of time . 2. 17. . 6. 18.3 av inst 0 0 ) f E ( d ) C % u E º C ð 7 ± % ` º % 14. s m/s m/s ( d ) g 1. 16. 6. 3. ' £ 7 2 E 2 C ð 8. ( ( d 2 7 2 Answers 4. 4. 6. 7. 39. ` 7 g g " 7 g " f t r 7 6.ANSWERS 265 15. " d 2 U d g ° 2 f % g 2 g d d ³ 2 ! 2 ` U E E 2 2 C C ð E " f 2 C ~ E g E f 2 2 d ( g " C f ~ " 2 C f E E 2 2 C C ð 1. 5. 6. m/s m/s m/s % U 7 ) ! E ) C ð 9. After s " £ ` 3.2 2 ` d 2 ` E 2 C ð ð ' d 2 g E 2 C ð ð Answers 4.001 5. m/s ( d 2 7 2 19. 2. then the quotient represents the approximate change in the cost per unit. (a) The quotient represents the average velocity during the time interval 2 2 { { 2 2 { { .01.1.2 m/s 7. during the time interval (c) If the number of units produced increases by .
3 2. 5. r E ) C ð 0 (b) The acceleration is decreasing.025% per hour. before 12 min and after 3. decreases on U E l 7 ' U f 2.99. " 7 ¤ ' r 4. ) ` % E ) C ð 2.5 E " 7 " f C E l 7 " C E " f 7 l f C 1.266 ANSWERS 15.6 and passing through the origin.] g " ! ! r 3. that person would probably have a hard time convincing the police that he or she was under the legal limit at the time of the accident. . . ' % E r % C ð U g % E U % C ð ! " % E ' % C ð g " % E g % C ð ` % % E " % C ð (a) (c) . (b) The graph is the straight line of slope 0. . Increases on C and . It is best to use central differences in this problem. m Answers 5.6 hours.08% per hour. Increases on . Approximately 0. approximately 0. decreases on E ' f 7 l f C E " g f 7 l f C E l 7 " g f C 3. Answers 5. . 7. % ' E C ð 0 ' ! ' E ` C ð 0 r r E ' C ð 0 (a) . The ball takes longer to fall than to rise.91. Increases on . The reading ) ' d § ) ' g d ° U ) E ) C ( ' d ) d § ' ) E ) C 6. decreases on g ~ ) º 9. we use central differences. g d ) U d ) d § ' ) E ) C 0 8. % % ¤ r % ¨ 7 ' ¤ 7 " ¤ 7 7 " % r 1. ' U ' % E % C 7 " % d § ) ' % E ) C (d) % u § ~ g f ð © 16. (c) 66 m/s Answers 5. The seventh observer has made an error.1 1. { Answers 5. . The equation of motion is should be 1. . [Use energy considerations. 2. 6.2 3. Note that if a person was involved in a serious accident after 5 minutes and then breathalysed by the police half an hour later. Again. Differentiating your answer should give the differential equation you started with. so is negative. not 1. The weight decreases as the ship moves away from the earth.
¥ ¦ 7 a ` a " " 7 r g ` a ' ` a ¥ 7 a " r % ¦ 7 ` g ` " 7 a % U a ' 1. 9. d Ù d Answers 6. 16. ÷ Ñ 14. 2. ± C ± ÷ 18. Increases on . " ÷ " d " ± g ÷ E " f ± g C ³ W ÷ ± 13.049. 15. 7. This is about 3% accuracy. U ® 2 % % ´ 2 ` " 7. 3/5 23. 9 20. 12. Use the fact that a " f is an arithmetic progression. 6 12. 7. height of sides: r " ` ' % " § ` % " ' 2 g f % 2 " g 2 ' f 9. d ¦ ¦ ¦ d ' d g d " 21. 22. 4 Ñ W Ñ r " W g " ´ " W U 9. global maximum of at cm r Answers 6. local minimum of % 2 ! f 5. local maximum of ' ® 2 " ! f % 2 6. 6. 8. f 2. Local maximum of at .ANSWERS 267 8. The rectangle should be a square of side cm. 3.3 ¦ ¦ ¦ 12. 4. " f E 2 f C E g d ± C ÷ f E " d " 17. Global minimum of at . " Ñ 9 7 " ± d ± 9 g " Ñ 9 7 g d " ± f g " 9 24. ' ¥ Ñ 9 7 ± g ± ` d " 9 ¥ " g Ñ 9 7 E g d ± ± ¥ C g 9 26. ` 3. 11. Width of base: cm. r ¤ f 10. no stationary points 10. No local extreme values. 9 " g 5. Global minimum of at . Local maximum of at % .1 ' " % f ¨ 1. ' % % % % % 7 ' a " U " E U % C y x w 13. 8. 10. 27. 5. ¨ Answers 6. 11. 3. 25. 19. decreases on at ( 4. global minima of at . no local maximum at E a g f 7 l f C E l 7 a g " f C 4. error 7 ' ` g ' 7 r 7 a ` " a % 7 % ` g " % 7 a ' r r U ! 6.2 § ' r " ` " 1. W .
10. 14. " g ' l 6. 1.6 ð ð 2. 8. 2 d 2 · E 2 C E C 7 2 d 2 E 2 C E C 5. 8. % ! ` ' 2. so . l l % Answers 6. . 11. 13. Ó g Ó ' f § ¥ 4. so it is differentiable for . . § § 9. 7. 5. by the ratio test. 7 " u 2 u l f u 2 u l f g u 2 u % l 3.4 12. . Similarly % d ð ð % d ð ð f ð ð ð 5. 4. The terms in the sum on the righthand side are smaller for a given value of . which is a geometric series with sum " u 2 u " f " u 2 u " f f Answers 6.5 1.7 3. 15. 10. 16. g f 2 " 3. 6. Radius of convergence § ð 4. 8. 2. " " a f ) g 6. so that we expect the righthand side to converge more quickly. E 2 d " C Ó E " d 2 d 2 g C g o " g g " E " d Ó C Ó 2 g d Ó 2 ' § Ó 2 g f o E Ó f Ó C ¥ ¦ ¦ E ~ " d C E ~ " f " C ` E " d ) g d ) C E ) U d E 2 d " C 2 g " E 2 d " C 2 U Ê ) ) U f ) d U É E ) C E C 7 E ) C E C ' d 2 E 2 C E C 7 E ' d 2 C E 2 C E C ' f " d 2 ' f 2 E 2 C E C 7 E 2 C E C Ý 2 d " E 2 C E C 7 E 2 d " C E 2 C E C E 2 d " C ~ " ¦ ¦ ¦ d 2 f ° 2 d 2 f " E 2 C ð ¦ ¦ ¦ d 2 ` " d 2 d 2 U d " E 2 C l % u 2 u g f l % " § ' " 7 ! r ! U ' " 268 14. 9. ' u 2 u a f ' g ~ " E ' % C y x w 1. 3. 2 Ó E Ó f " C g " § § 7. 9. ³ § § Answers 6. . ¥ ) ' C U 1. From the previous question. ANSWERS error ¥ Answers 6. Ú Ú 10. 5. 4. The series converges for We then have . r E 2 C E C 7 r g " E 2 C E C 7. .
Here is one of many possibilities: % u 2 u l f 7 " ¯ a g ¯ g d ¦ ¦ ¦ d d r d " 11. 7. Answers 6. ¤ 15.1 ¤ 2 Õ 2 U f 10. ¤ 2 ó Î ò 1. L as 2. L terms) does not exist.8 2. cosec 2 U . ¤ ¤ 9. 12. NL ( term) { 7. 6. y § Answers 7. 3. 18. ¤ 14. L cosec kg ANSWERS 269 . f 2 ó Î ò f 12. 5. ¤ ¤ 11. The displacement after 2 s is 6. 16. cosec C m. .Ó d Ó { Ó E 2 d C { ° ° ¥ £ £ Ó f Ó g { Ó g f r { § j U § Ù T Ó d T j Ó U § Ù { ¥ ¥ Ú ¥ £ Ó d Ó { § ¥ £ Ó E 2 d C { Ó E 2 d C { ¥ ¥ £ £ Ó d Ó { d Ó { ³ ¥ ¥ ¥ £ £ % ¤ 2 " 2 . L 3. M ò 2 g 8. L 5. " 11. 8.6 õ ô E 2 C 6.3 % g Answers 7. NL ( ð { { 4. ) ' ó Î ò ) g ó Î ò ' f ) ' ò õ ô ) g ò õ ô g 9. õ ô 2 2 g f 2 2 ô 2 g ò õ ô 2 ó . 7. 2 ó Î ò E 2 C 7 2 ò Answers 7. term) 10. 7. 2 2 ð 2 E ð ð 2 C Ô 9. NL ( . mg 4. g d 2 g ó Î ò 2 ô M ò E % C t % k 2 7 2 ò õ ô E 2 C 2 ò õ ô ' 2 ò õ ô f " E 2 ó Î ò d " C 2 ò õ ô 2 ó Î ò ` f 2 ó Î ò 2 f 2 ò õ ô 2 g ó Î ò Ó f ò õ ô Ó g 2 ò õ ô 2 ó Î ò g % g g ò õ ô r % % ' ' n U " U ' Ó U { r ! U " ` % l u 2 u % 7 " f E 2 C v 12. 8. õ ô Answers 8. ´ r r 1. ð ð 13. ' ¥ 17. ¤ ¤ 13. 10.2 U % % % ` 2. 8.
" r ' Answers 8. 3. ¤ 3. 2 12. 15. 4. £ 38. 32. 28. 30. 10. 18. 16.3 g g 2 ® ® ' ® " f C U g g ' 29. cis 26. 4. ¤ 26. " 33. 31. 2 2. 6. ¤ 22. ' " 1. 23. ' f { 5.2 j Ó 1. ¤ Answers 8. 7. ¤ ¤ 31. 19. 14. ¤ ¤ 39. ¤ £ ¥ Ó d Ó { ¥ ¤ 21. ANSWERS 20. cis . % " ~ " f 17. 24. 30. ¤ ¤ 29. cis " " 2 d 21. 20. ¥ 1. ¤ ¦ ¦ j ³ Ù Ó d j Ó ³ Ù { Ú ¥ ¥ ¥ ¤ 33. ¤ 23. U g 25. ¤ ¥ £ j Ó E 2 d C { § ¥ ¤ 25. r f 9. ¤ 35.4 Ó § Ù § Answers 8. ¤ 34. ¤ 32. 2 r " f r f 13. cis 27. ¤ ¤ 37. 11. 8. £ ¤ 40. 2. cis 28.E 2 ' ó Î ò d 2 ' ò õ ô C Ó { 2 ó Î ò d 2 ò õ ô { £ ¤ £ g g ó Î ò d ò õ ô j Ó { ¥ ¤ 2 ' 2 ' £ E 2 g ó Î ò d 2 g ò õ ô C Ó { ¥ £ g g g g 2 ® f 7 " 2 ® 7 " f E g 2 2 ® 7 " ® ' " ' " g U g E r " 2 ® " f C g ' ` U ' r ~ % " U g ³ ³ 2 f § 2 § Ù % 2 a d " f 2 ! f g r g r E 2 U d ' C % " f % " d r ' " 2 g r Ó d j Ó ü § ´ j Ó ) d ¥ ¥ £ ¦ ¦ ¦ ¦ j Ó d j Ó { Ó d Ó { § § ¥ ¥ ¥ £ ¦ ¦ ¦ ¦ Ó d Ó { Ó ) d Ó { § § § £ £ ¦ ¦ ¦ ¦ j ³ Ù § Ó d j Ó ³ Ù § j Ó ) d j Ó 2 ³ ³ ¥ Ú ¥ ¥ £ £ ¦ ¦ Ó d j Ó 2 ° ¥ £ £ ¦ ¦ ¦ ¦ Ó ) d Ó { Ó d Ó { ° ° ¥ ¥ ¥ £ £ ¦ ¦ ¦ ¦ j Ó E 2 d 2 d C { Ó d Ó d { § § ¥ ¥ ¥ £ ¥ £ Ó d Ó E 2 d C { Ó d Ó E 2 d C { ¥ ¥ ¥ £ Ó d j Ó { § ¥ £ £ j Ó E 2 d C d { Ó d Ó E 2 d C { § ¥ ¥ ¥ £ j Ó E 2 d C { ¥ £ £ 270 19. 36. cis 24. ¤ 27. f f 2 d 22.
r 3. ¤ 7.5 õ ô g f { 2. Guitar string. ¤ 5.g f Ó d Ó { ¥ £ § 2 ó Î ò d 2 ò õ ô § f Ó d Ó { 2 g ó Î ò f 2 ó Î ò d 2 ò õ ô { ¥ £ E ' a " d ) a g U U C ó % " U " E r ~ % " ) a C ó Î ò f E r ~ % " ) a C ò õ ô " " ` t E d ) " " ò õ ô C Ó 2 ³ ¥ ) " " ó Î ò r ¦ r t E ) r d " C Ó 2 ³ ¥ § ¦ U t E Ó U f Ó C f 2 ° µ ¦ ¦ § r 9 t E Ó U f Ó C f 2 µ ¥ ¥ ¦ ¦ U 9 t Ó ) U d Ó 2 ° ° ¥ ¥ ¦ ¦ ' 9 t ) a ó Î ò E a ~ ' C d ) a ò õ ô C Ó 2 § ¦ % g ' ± t § g ) ò õ ô 2 " w a U " ~ ) ) U " ó Î ò { g E g ~ ) ò õ ô f g ~ ) ó Î ò C j Ó { E ) U ó Î ò f ) U ò õ ô g { ¦ E ) ` ó Î ò d ) ` ò õ ô C Ó { E ) ó Î ò d ) ò õ ô C Ó d Ó { ¥ £ £ ¥ ¦ ¦ ¦ g g E ) ó Î ò d ) ò õ ô C j Ó d Ó { ¥ ¤ ' ' £ ¥ ¦ ¦ ' ' E ) ó Î ò d ) ò õ ô C j Ó { § ¥ £ ¦ E ) ó Î ò d ) ò õ ô C Ó { ) ' ó Î ò d ) ' ò õ ô { £ ¤ £ ¦ Ó d E 2 ó Î ò d 2 ò õ ô C Ó { E 2 g ó Î ò d 2 g ò õ ô C Ó { ¥ ¥ £ £ E 2 ' ó Î ò d 2 ' ò õ ô C Ó { 2 g ó Î ò d 2 g ò õ ô { ¥ £ £ 6. tides. midday position of the sun as the seasons change. alternating electric current.6 12. ¤ £ 1. ¤ ¤ 9. 13. pendulums. 6. ) ò Answers 8. 10. 12. ¤ 11. Î ò ' % " % n % " E ) C 2 " 11. ¥ 9. N/m cm. or sec kg/m 16. ¥ ¥ 10. ¤ ¤ 15. After 1. " 17.369 sec 2. g ¤ g 13. 2 ó Î ò ¤ ANSWERS 271 . 14. 5. 8. frequency kg s ¥ 8. Shock absorbers. " Answers 8. resistance in an AC circuit. U u r 7. ¤ 3. N sec/m or kg sec/m E " ! ' g C ' ' ' ' g 7 E " U ! " ± C 7 U U U U " ± 4.
Ó ' d ) ó Î ò d ) ¦ ò õ ô £ E ) C 2 ¦ a ' " d ¦ § Ó ¤ d ¦ Ó £ E ) C 2 10. 2 ò õ ô 2 g f 2 £ ó Î ò ' d 2 ò õ ô £ g { 7. ¦ ) g ¦ d ¤ d § ¦ ¥ Ó £ E ) C 2 § Ó ' ¦ d Ó ¤ d ¦ Ó £ E ) C 2 12. ¤ 15. ) ó Î ò ) " g d ) ó Î ò ¤ d ) ò õ ô 2 6.272 ANSWERS 18. F ' d ) d § ) 7. ¥ Ó ) r f ¥ Ó d ¥ Ó E ) C 2 ¦ ¦ " ' £ f ¥ Ó ¤ U d " % " § Ó E ) C ( E ) C 2 14. . Regular pushing on a swing. £ . ) ò õ ô ) g f ) ó Î ò £ g { 8. Unwanted feedback in public address systems. (a) . 13. ) ' ó Î ò ¥ Ó E ) C 2 t ) ' ó Î ò U d ) ' ò õ ô g É ¥ Ó f g E ) C 19. ð 20. The circuit resonates with the transmitted signals and absorbs peak energy from the signal. ¤ 11. 2. the natural frequency in the receiving circuit is made equal to the frequency of the broadcast waves of the sending station. The initial value problem describing the motion is . ¦ £ E ) ò õ ô f ) ¦ ò õ ô C ¦ f £ { 9. Tuning a radio: by tuning a dial. The solution to this is U d ) r ó Î ò U f E ) C 2 % E % C ð 2 g l E ) C ) 2 g t ` E E £ ) ) C g 2 ó Î ò 7 d g ) r g Ê " d ò ' õ ¥ ô Ó r ¤ f ' ¦ ¥ ¦ d § Ó ¥ ° ¤ Ó ¥ ! Ó d ¦ f ° ¦ ¥ Ó ¥ E ¦ Ó £ ) ' C " 2 C E 16. 2 g ò õ ô a g " " f 2 g ó Î ò a ' " d 2 g ó Î ò ¥ Ó ¤ d 2 g ò õ ô ¥ Ó £ { 5. % E % C 2 ) g ó Î ò % E g " f ¦ Ó 2 C % ¦ r § a ¥ Ó d £ E ) ) g C ð ð ó ¦ E Î ) 2 ò C % 2 ' ' 17. ¤ d ' ~ § E g f 2 C ¤ d ' ~ § 2 f U ~ ° 2 5. 2 r g ' ó Î ò ¤ d 2 r g ' ò õ ô j Ó d 2 £ ' ò õ ô 4. 4. " 2 g f ¤ d 2 ' ò õ ô § f 2 ó 3. (a) (b) % % % " ( Answers 9. 6. The transient part of the .3 ¤ d 2 g ò õ ô f U ~ ° Ó ¤ d 2 d 2 d U ~ ° 2 1. (b) solution is ) C 2 g r so as and the steady state part is . .
The area of the crosssection is 11. 11. 42 Answers 9. 152 7. 28 14. 2 ó " 3. 19 intervals 10. 20 intervals Answers 9. 5. ° 2 2 g f " 2 g 7. t ° § 2. g ¦ 3.1 m . the volume of water per second is 5. E 2 d " C 2 " § d 2 g 2 d E " d 2 C ó Í 2 g 5. . f " 8. d " 10. 2 12. d 2 g C 2 § Ó d 4. 124 16. U r g n ¥ Ó Ó d f " 11. (In fact 2.9 2 g g 1. 6. ` 9. ) 2.5 2. 8. Í 2 2 " d 2 E § 2 C 2 ' E ° 2 d U 2 g d C 2 2 g g 7. 125 15. 2. U 7 ` ! r U % ¤ t r a ! a ` ` % ` ! ¤ 9 ¤ 9 7 ! 7 g r ! a a ! % ¤ ¤ t 1. 8. 2 g d " 6. Yes Answers 9.6 1. 518 E a d 1 ' d 1 g C 1 " § r a a ` ! " ` r ° ` g U ü 2 " U " % ~ f ó " Î ò f ¤ ¤ ü ` E ¤ 2 U 9 9 ' " 8. d " 4. 155 4.6498 (6 subintervals) 9. 10.512) (b) A straight line f f 2 g 2 C E d 2 g 2 a d C " g C § 2 2 d ' " d E ¥ 2 2 g 9. (a) . t 7. 686 5. E " Answers 9. 259 E " d ' f g C " 3.ANSWERS 273 Answers 9. 44000 7 7 2.51 m /sec. 506 6.7 3. 9 13. 6.8 2 g ó Î ò d " 2 d " 1. 2. 9 7 r r " f ¤ t ' ! a U g ¤ 9 7 r a a U g ¤ 5. 155 12. 7 g ! " g " ¤ t 4.
18. Newton’s rule can be used to compute function values. 2.346574 2. 1. 7. 25. (c) About 20 000 kilograms 2. 8.1 ' f 2 E 2 C ¥ 1.25992 0. has two solutions for   6.523599 14.08 pm E 2 ó . 28. 10. 15. The population doubles after 25 years and triples after 39. 19.41421 1. 16. E 9. It is difﬁcult to ﬁnd a formula for the inverse. Answers 10.739085 3. 21. 10. 3. 6.44949 0. About 14 minutes ° % " § ! r ' ± 26.0855 hrs ¥ ¥ ' n g ö õ Í % " 13. (b) 14. g r f 2 E 2 C ¥ is not 1–1 8. 20. 3.2 2 2 2 1. 4. (a) 10 Watts/m . The function is not 1–1.86 days 29. 3. 2 ~ " d 2 4.274 ANSWERS 15. The function is not 1–1.86081 1. 23. 12. 10. . E 2 C ¥ " f " É E 2 C ¥ 5. " 2 d " 2 ~ Ó E 2 g ö õ Í d 2 ö õ 2 Í ö 2 f õ " Í " C d g Ù 2 ö E E 2 2 ö õ ` 2 2 Í C õ ö Í ö g C g 2 õ õ Í Ù ¥ ~ Í C g " 11.25992 0. 11. 22. 7. 0.876726 1. (b) 130 Db Answers 10. This can be proved by showing that .817 days. 2 f 2 ò õ ô ö õ Í C E 2 ò õ ô C E 2 2 ó Î ò d 2 ö õ Í ò õ ô C Ý Ü Û 2 8. 37 000 years 30. 17.6 years. 5. § j 2 f " E 2 C ¥ § j E " d 2 C E 2 C ¥ 9. 2 1.44949 27. The function is 1–1 and so has an inverse. " " d f 2 2 ' § j Ê § g 2 % " § ` g " " w 13. Take day .47502 20.
¤ ° 2 d " Í d " C 2 g 1. ð 12. 2 ó Î ò C D 14. " " 23. ð ð 17. 7. 8. ð 15. 13. 21. ð ð 19. 5.2 C Ó E ) C 2 13. 6. ¤ 3. 4. 18. 24. ó Î ò d " ¨ 2 ò õ ô { ¨ 27. 20. 26. 9. 20. 23. 2 " 2 17. 16.ÿ g þ 2 r D F õ ô r E 2 C E 2 C ð 2 D ò õ ô f " E 2 C ð 2 2 D ò õ ô f E 2 C ð 2 D ó F f E 2 C ð 2 E 2 ö õ Í C D ó Î ò E 2 C 2 D ó Î ò 2 ' E 2 C ð § ð 2 D ó F E 2 C 2 D ó Î ò 2 D ò õ ô ' E 2 C 2 g D F õ ô g E 2 C 2 g E 2 C 2 D ò õ ô 2 ' E 2 C E " d 2 g C D ó Î ò g E 2 C § ð E ) ' D ó Î ò d ) ' D ò õ ô £ ¦ E ) g D ó Î ò d ) g D ò õ ô C Ó E ) C { § ¥ £ ¦ ) g D ó Î ò d ) g D ò õ ô E ) C { ) ' D ó Î ò d ) ' D ò õ ô E ) C { £ £ 2 f 2 ¨ f "   E C ð j § 2 E 2 f U C 2 ¨ 2 2 d " Ê f 2 ò õ ô É Ó E 2 C ð ¥ · " d 2 E " f 2 C f " E 2 C ð E 2 C ð 2 2 2 2 2 2 d " 2 d " 2 f " · g ~ " U ~ f g ~ g ~ g g ~ ' U ~ U ~ f ' ~ ' ~ g ~ g ~ Answers 10.1 11. E 2 ö õ Answers 11. 15. 10. 18. E " 26. ¤ Answers 11. 2 f " 22. csch 16. 2. " f 2 g f 24. " 21. 2 { ¨ 28. ¤ 4. 22. 0 2. " 25.5 1. 3. 2 f ' { ¨ 29. ANSWERS 275 . sech 2 ò õ ô 2 ò õ ô E 2 C ð 25. 19. 14.
9. 3. 93 8. 2 g ± E ) C º 1. 12. 10. 1 2. 10.3 Note: The constant of integration has been omitted in the answers below. 14. 3. ' ¥ g E 2 g D ó Î ò d " C ö õ Í " 15. " " " 5.2  2  E 2 C ð 1. 4. Answers 12. 2 6. ü ó Î ò ü f " ü 17. 21/64 g ~ 5. § 2 ó Î ò 10. E 2 ö õ Í C " 13. " 9. 8. " 18. 9. g g g E " d 2 C 4. 0 6. . 2 2 Answers 12. sech 4. 2. 16.3 ANSWERS 2. § § § 7. No matter how far it falls. C E 2 ó F 19. it never exceeds this velocity. E 2 D ó Î ò C ' 2 D ó Î ò r 7. g 3. g Ó f " · D ó F f E 2 ó Î ò C ó Î ò ¥ g g E 2 ö õ Í C ö õ Í ¥ g g g d f E ) ö õ Í C  ü D ó Î ò  ö õ Í g ' Û Ó ê Ó f j E ¤ d º g d º r C § © ' 2 ò õ ô g f ) d r g ¤ ' " d 2 · j E U d 2 C j E " d 2 C U Ó o f " f ' U D ó Î ò U ~ f g U ~ " d 2  2  " f 2 E 2 C ð 2 D ó F E 2 f " C ¥ 2 E 2 C ð E 2 C ð 2 d " " d 2 D ò õ ô r · E 2 C ð E 2 C ð ¥ " f 2 " d 2 2 D ó F E 2 f " C ¥ d E 2 C ð f E 2 C ð · " f 2 U E 2 d " C d " E 2 C ð E 2 C ð º d ) D ò õ ô ö õ Í Ý · 276 27. (b) The limiting velocity is . 1. 11. This is the terminal velocity of the body. 4 7. 11. 2 ° ò õ ô 5. 2 8. ± ~ (c) ± Answers 11. 6. 20. ¥ 12.
¥ ¥ 12. 4. 2. f ¥ g g E 2 d ' C ö õ Í f E 2 d g C ö õ Í g d E 2 d " C ö õ Í f g Ê " d 2 ` ' Ê g f 2 ' É ö õ Í d ó . d f ¥ ' ' g Ê E " d 2 d 2 C ö õ Í f 2 ö õ Í d É ó . § 2 Answers 12.5 a g § 2 g § 2 g 1. 2 d " " 2 d " 3. 14. " f 2 " " Ê " d 2 g É ö õ Í g f 2 5. " f 2 " 2 g 11. ¥ 5. f f . ' E 2 ö õ Í C d 2 ö õ Í f 2 2 f f E 2 ó Î ò C 2 d 2 ó Î ò 2 f " · g d 2 g f Ê g g 2 É . 10. " 2 " 7. 10. ¥ Ê g d 2 r É ö õ Í Ê " f 2 g Ê 2 d g É ö õ Í É ö õ Í 2 ó 2 f " · E 2 d g C d 2 ó Î ò 2 Note: The constant of integration has been omitted in the answers below. 8. 2 ¥ ¥ f " · d 2 ó Î ò 2 2 ó Î ò 2 g d 2 ò õ ô E 2 f g C 2 ó Î ò d 2 ò õ ô 2 f Ó f Ó 2 Note: The constant of integration has been omitted in the answers below. Answers 12. 2 d " g g E 2 d g C ö õ Í d E 2 d " C ö õ Í g f r 2 ö õ Í 7. 6. 3. 4. 2 ö õ Í " 13. É ö õ Í ¥ 2 d " Ê É ö õ Í d 2 ó . ANSWERS 277 . ' f 2 " 6. ó . ¥ Ê " d 2 U g É ö õ Í d 2 ó .4 1. 2 d E 2 ò õ ô C ö õ Í ¥ ' r 2 ö õ Í d f E 2 g ó Î ò f 2 g ò õ ô g C f ' g E 2 d " C ö õ Í f 2 ó . cosec 2 ò õ ô d  2 . 2 d U " 2 d U 11. õ ô f 2  ö õ Í d E E 2 ó Î ò C ö õ Í C ò õ ô f 14. § 2 § 2 Ó 9. 16.j ³ § j ü Ê d " Ê " d ) É ö õ Í r ) d É ö õ Í ' d j § ) ` j ³ ü j " f § ) Ê 2 ò õ ô d g U É ö õ Í g ó . r f 2 a g E 2 d " C ö õ Í f 2 ö õ Í " 9. 2. 8. 2 ò õ ô d " 15. " 2 g d " " 13. ' " 12.
d 7 2. C d C 10. ó . B ¥ C Ê d d + ö õ Í d É ó . ¥  = d C + ö õ Í d  B C + ö õ Í C Ê d d + ö õ Í d É ó . = É ¥ Ê C C É ó . 8.7 278 1. Answers 12. d C + ó . C d 3. B ¥ Ê C C É ó . B ¥ C  H d 7 d + ö õ Í d  d C + ó . C d 9. Ê ö õ Í d É ó . 4. ¥ ANSWERS . B ó Î ò 6. ¥ Note: The constant of integration has been omitted in the answers below. d 7. ¥ Ê Ê C H d < d + ö õ Í C d É ó . Ó C d C 5. ¥ Ê É ó .
116 of . 146 antiderivatives. 72 concave up. such as a deﬁnition or an explanation. 159 Cardano. 138 polar form. 135 series. 52. 4 asymptotes. 139 continuity. 125 critical damping. 76 chain rule. 28 at a point. 73 conjugate pairs. 7. 130 repeated roots. 45 of . 61. 135 modulus.Index Bold face page numbers are used to indicate pages with the primary information about the entry. 138 of SHM. 55 alternating series. 138 Aristotle. 145 de Moivre’s theorem. . 169 Anzac Bridge. abuse of notation. 28 continuous function. 9. 33. 68. 138 imaginary part. 72 concavity. 132 backward differences. 125 of . 143 decreasing function. 188 of . 32 convergent series. 204 of a constant function. 138 exponential. 82 cosecant. 37 on an open interval. 6 auxiliary equation. 138 phase. 49 ò ó ö ó ¤ õ Î õ ô ò Í . 134 amplitude. 149 damping. 96 complementary equation. 58 acceleration. 61 barometer. 5 beats. 69 derivative. 153 complex conjugate. 138 real part. 23 atmospheric pressure. 134 central differences. 113 cotangent. 135 composition of functions. 96 concave down. 141 plane. 116 of . 37 continuous. 138 argument. 153 function. 109 Archimedes. 125 cosine function. 133 distinct roots. 171 Argand. 114 amplitude of a complex number. 138 Cartesian form. 136 Argand diagram. 176 area under a curve. 37 convergent sequence. 140 complex number. 33 on a closed interval. 129 complex roots. 135 argument of a complex number.
17 plotting with Mathematica. 44 differential equation. 25. 69 primitive. 230 inverse sine. 12 domain. 256 kinetic energy. 205 inverse hyperbolic functions. 75 homogeneous. 141 Euler’s formula. 61 central. 15 interval closed. 218 inverse trigonometric functions. 114 Euler. 254 energy. 239 intersection. 218 Jacobi amplitude. calculating. 103. 57 smooth. 12. 69 differentiable. 204 of a sum. natural. 170 indeﬁnite. 130 geometric series. 93 differentiation by ﬁrst principles. 50 of a quotient. 76 forward. 15 inverse function. 129 linear. 17 hectopascal. 200. 61 homogeneous linear. 12 continuous. 46 ﬂight 811. 169 rational. 12 Galileo. 70 ﬁnite differences. 32. 141 exponential function. 17 increasing. 61 differentiable. 197. 28 domain. 215 lefthand sum. 37 decreasing. 129 second order. 3 elementary function. 68. 66 general exponential function. 252 elliptic functions. 44 graph of. 61. 52 of a square root. 82 graph of a function. 49 of an inverse function. 61 frictional forces. 33. 204 of inverse trigonometric functions. 241 integration by substitution. 193 discontinuous. 111 solved using Mathematica. 11. 81 initial value problem. 15 open. 225. 12 integral. 201 inverse function values. 156 forward differences. 45 logarithmic. 227 increasing function. 15 inﬁnite. 67 differentiating a power series. 5 forced oscillations. 177 Leibniz. 69 indeﬁnite integral. 215 error term. 28 value. 63 input number. 212 general solution. 15 intervals. 41. 130 . 129 hyperbolic functions. 58. 170 integration by parts. 12 inﬁnite series.282 INDEX of a product. 170 independent variable. 45 linear combination. 13 Dreamworld. 28. 212 extreme value. 219 differences backward. 128 function. 27. 225 elementary functions. 8. 4 ﬂuid pressure. 129 nonlinear. 87 general.
8 Olympic Games. 45 sequence. 90 range. 206 rate of change. 70 logarithmic differentiation. 251 periodic.simple. 21 potential energy. 113 solution general. 125 second derivative. 32 simple harmonic motion. 170 primitive function. 66 point of inﬂexion. 103. 116 phase of a complex number. 32 divergent. 138 polynomial. 161 stretched string. 122 periodic functions. 130 of a differential equation. 54 ratio test. 141 rational function. 73 horizontal. 241 pascal. 62 square wave. 127 strong damping. 153 particular solution. 176 . 155 millibars. 70 local minimum. 70 method of undetermined coefﬁcients. 13 natural domain of a rule. 146 simple pendulum. 12 Riemann sum. 206 nonlinear. 25 rational functions. 12 partial fractions. 215 primitive. 148 substitution. 206 Newton–Raphson method. integration by.INDEX 283 differential equation. 239 sum Riemann. 251 Simpson. 57. 12 Raphson. 103. 71 local. 177 secant. 243 partial sums. 193 maximum absolute. 11. 7 quadratic integrals. 50 projectile motion. 145. 32 convergent. 82 particular integral. 116 periodicity. 50 radius of convergence. 244 recurrence relation. 71. 1 one–to–one. 130 products of functions. 146 physical laws. 176 righthand sum. 247 quotients of functions. 85. 71 global. 71 local. 71 global. 252 numerical properties. 128 local maximum. 61 particular. 158 restriction. 103 trial. 153 parts. 169 principle of superposition. 75 minimum absolute. 138 phase angle. 75 pendulum. 71 steady state. 183 sine function. 138 natural domain of a function. 206 Simpson’s rule. 70 modulus of a complex number. 12 Newton. 200 output number. 112 resonance. integration by. 45 Newton’s method. 71 polar form of a complex number. 58 pump. 30 stationary point.
49 superposition. 7 tangent. 154 union. 104 trigonometric functions as ratios. 116 turning point. 71 undetermined coefﬁcients. 104 velocity. 53 weak damping. 57 transient. 147 колхоз 9:02 pm. 182 trial solution. 62. 119 trigonometric identities. principle of. 10/4/05 . 15 variable.284 INDEX sums of functions. 5 Tower of Terror. 125 Torricelli. 1. 160 trapezoidal rule. 9 variation of parameters. 130 Tacoma Narrows Bridge.