You are on page 1of 21
1 Basic Concepts In many branches of mathematics, itis convenient to have available a notion of distance between elements of an abstract set. For example, the proofs of some of the theorems in real analysis or analytic function theory depend anly on a few proper- ties ofthe distance between points and not on the fact that the points are in R or C. ‘When these properties of distance are abstracted, they lead to the concept of a metric space. The notion of distance between points of an abstract set leads naturally to the discussion of convergence of sequencss and Cauchy sequences in the set. Unlike the situation of real numbers, where each Cauchy sequence is convergent, there ate metic spaces in which Cauchy sequences fal 1 Louverge. A ‘metric space in which every Cauchy sequence converges is called a “complete” ‘metric space. This property plays a vital role in analysis when one wishes to make an existence statement. ‘Our objective in this chapter isto define a metre space and lista large number of| examples to emphasise the usefulness and the unifying foree ofthe concept. We also define complete metric spaces, give several examples and describe their elementary properties. In Section 1.5, we shall prove that every meiric space can be ‘completed’ in an appropriate sense. 1.1. Inequalities ‘The subject of inequalities has applications in every part of mathematics, and the study of metric spaces is no exception. In fact, the definition of a metric space involves an inequality which is a generalisation of the familiar triangle inequality, satisfied by the distance function in R. (|x — y| =| —2| + |z — yl forall x. zinR orc) Isis seution, we establishe sone inequalities Uuat will be requited for cout "8 that some of the examples we list are indeed metric speces. Theses examples will be invoked repeatedly Proposition 1.1.1, The function f(x) 0, is monotonically increasing, 4 1. Basie Concepts Proof. Let y > x = 0. Then 1 1 dst Tey “Tre Tey” | Tee“ Try Tre ‘Theorem 1.1.2. For any two real numbers x and y; the following inequality holds: T+ kk+yl byl T+ )e+ yh T+ Proof. Let x and y have the same sign. Without loss of generality, we may assume that x=0 and y2=0, and so. ety] __ ety y Telet yl eety res] «42 Yb Tox Tsp T+hl T+ Suppose x and y have different signs. We may assume that |x| > |y|. Then bets It follows from Proposition 1.1.1 that ety lal l+|ety| 14+/ ‘This completes the proof. A ‘The next proposition is the well known arithmetic mean-geometric mean inequality, or AM-GM inequality, for short. Proposition 1.1.3. (AM-GM Inequality) If a> 0 and 6 >0 and if 0 0, i concave, we have Ina + (1 A)b) = Alna + (1A) IMB, In@b* = In(ha + (1 ~ 95). exp xis astrictly increasing function, it follows from the above inequality that b= ha + (1 —dYb. a Remark. When x20, y=0, p> Land jp + fq = 1, we have 1.1, Inequalities 25 Lyd ysis siy ‘The inequality is obvious when either x or yis 0. Ix # 0-4 y, then it fllows from Proposition 1.1.3 upon writing A = /p, a = x and b-* ‘Theorem 1.1.4, (Holder's Inequality) Let x)=0 and y,=0 for i= 1,2,...4m and suppose that p > 1 and q > 1 are such that 1/p + 1 = 1. Then In the special case when p = q = 2, the above inequality reduces to 1a fer() ym Proof. We need consider only the case when So af 0% So yf. To begin with, we assume that a a y af 3 af (4) In this case, the inequality (1.2) reduces tothe form ‘To obtain (1.5), we put successively x= xj and y= yj for i= 1,2,...,m im the inequality of the preceding Remark and then add up the inequalities so obtained, "The general case can he recced to the foregoing special case if we take in place of the numbers x, the numbers for which the condition (1.4) is satisfied, It follows by what we have proved in the paragraph above that Duy ‘This is equivalent to (1.2). This completes the proof, o 6 1. Basie Concepts ‘Theorem 1.1.5. (Minkowski’s Inequality) Let xj=0 and yi =0 for i= and suppose that p= 1. Then s2aeveatt Mp (Serr) “=(e)"Ey) 6) Proof. If p = 1, the inequality (1.6) is self-evident. So, assume that p > 1, We write Dh Ga + y)? in the form Doty =< Toasty + Does tye aa Let q > 1 be such that 1/p + 1/4 = 1. Apply Holder's inequality tothe two sums on the right side of (1.7) and obtain Lewtns (Ss) "(Suene)) (64) Earn) EYE)" TE~»)" iding both sides of this inequality by (377, (x; + y,)")'%, we obtain (1.6) in the case 52" (ay + y)? # 0- In the contrary case, (1.6) is self-evident. a ‘Theorem 1.1.6. (Minkowski’s Inequality for Infinite Sums) Suppose that p=1 and let {Sely=n [ahr=1 be sequences of nonnegative terms such that 37 4" and yi? are convergent. Then So (%y + Jn)” is convergent. Moreover, Broo Foray pst E+») "=(E<) “(Ee nteger m, we have from Theorem 1.1.5, £6) Be" 1.2, Metric Spaces ” ‘Thus, (37%, (x4 + y,)°)"?, which is an increasing sequence of nonnegative real ‘numbers s bounded above by the sum x \YP pm \ MP (=) +(&») Itfollowsthat > (a + ya)? isconvergentand thatthe desired inequality holds. (1 ‘Theorem 1.1.7. Let p> 1. For a= 0 and b=0, we have (a+ oy sZt "Car + OF, Proof. If ether a or bis 0, the result is obvious. So assume a > O and b> 0. Now the function that maps every positive number x into 2” is convex when p > 1. So, a+b)? w+ z 7 (a4 bP Sah +B, o 1.2. Metric Spaces ‘The notion of function, the concept of limit and the rated concept of continuity play an important role in the study of mathematical cnalysis. The notion of limit can be formulated entirely in terms of distance. For example, a sequence {xhy21 of real numbers converges to xif and only if forall ¢ > 0 there exists a positive integer nm such that [sp — x] < whenever n= 1, A discerning reader will note that the above definition of convergence depends only on the properties of the distance |e —B| between pairs c, of real numbers, and that the algebraic properties of real ‘numbers have no bearing om it, except insofar as they determine properties of the distance such as, Ja —B| > O when a # B, la —B| = |B — a and a—y1= ~ Bl + |B — >I. ‘There are many other sets of elements for which “distance between pairs of clements” can be defined, and doing so provides a general setting in which the notions of convergence and continuity can be studied. Such a setting is called a ‘metric space. The approach through metric spaces illuminates many of the concepts of classical analysis and economises the intellectual effort involved in learning them. ‘We begin with the definition of a metric space. Definition 1.2.1. A nonempty set X with a map d:X x X — Ris called a metric space if the map d has the following properties: 28 1. Basie Concepts (MSI) diuy20 my yeX (M82) d(x.y) = 0 ifand only fx = ys (M83) d(xy)=dQ.2) uy EX (MSH) dlx,y) SaG52)+dey) sy, ex. ‘The map dis called the metric on X or sometimes the distance function on X. The phrase (Xd) is a meteic space” uveans Uhat lis a uetsic om the set X. Property (MSA) is often called the triangle inequality. ‘The four properties (MSI)=(MS4) are abstracted from the familiar properties of distance between points in physical space. Itis customary to refer to elements of any inet space as poin ) as the distance between the points « aud j. ‘We shall often omit all mention of the metric d and write “the metric space X” instead of “the metric space (X, 4)". This abuse of language is unlikely to cause any confusion, Different choices of metrics on the same set X give rise to different metric spaces. In such a situation, careful distinction between them must be ‘maintained. Suppose that (X,d) is a metric space and Y is a nonempty subset of X. The restriction dy of d to ¥ x ¥ will serve as a metric for ¥, as it clearly satisfies the metric space axioms (MS1)-(MS4); so (Ys dy) isa metric space, By abuse of language, ‘we shal often write (Yd) instead of (Y, dr). This metric space is called a subspace of X or of (Xd) and the restriction dy is called the metricinduced by d on ¥. Examples 1.2.2. ‘metric on R, the set of real numbers. To prove that only (MS4), as the other axioms are obviously sai ‘The function d:R x R—+ R* defined by d(x, ) = |x — yl isa isa metric on R, we need verify fed. For any xy,2 € Ry (x,y) + dlys2). d{x,2) = [x2 = xy) + -D1S |x y+ It is known as the usual or standard metric on R. (ii) Let = (Sisd2s-se93):8) €R, LSS} be the set of real tuples, For x= (3%55.-+ 4%) and y= (jisJoy--sy%) im RY define dey = (Saw) (For m= 2, dlxy) = (x) — y)* + (22 —y2)°)"? is the usual distance in the Car- tesian plane.) To verify that d is a metric on R®, we need only check (MS) i, if isZb---+2)s We must show that dlxy)=d(x2)+d(z,y). For S12... set = 2 bk = 2 — Yh Then dlx,2) and 1.2. Metric Spaces 2» (xy) = (Se + wr) : We must show that (Sa + wr) = (= 4) ‘ (5 8) o a & a Squaring both sides of (*), and using the equality (a+b = 2+ 20b+B, ‘we see that (*) is equivalent to which is just the Cauchy-Schwarz inequality (See Theorem 1.1.4). This metre is ‘known as the Euclidean metric on R", When > 1, RY! can be regarded asa subset of R"in the usual way. The metric induced on R""! by the Euclidean metric of R" is the Euclidean metric of R™!, (iil) Let X = RY, For x = (8183p. 005%) and y= (jt Yars Je) im RY define 4a = (& bs or) ‘ where p= 1. Note that when p = 2 this agrees with the Euclidean metric. To verify that dp is a metric, we need only check that do(x.y) = dy(x.2) + dy(zsy) when 25 (2is2is.04%) ERY For k= 1,2,...9m Set te = ae — 2, be = a — Then (Emr) Slat wr)" y(o2) = (Siar), ley) and aly) = ( ‘We need to show that (Sia sae)" (Siar) “(Su)" However, ths is just Minkowski’s inequality (see Theorem 1.1.5), 30 1. Basie Concepts Here again, when 11> 1, the metric induced by d, on the subset R*™? is the corresponding metric of the same kind, ie. et 7 dy(xy) = (E» -»t) when 4y € R™! (iv) Let X= For x = (ser) and Y= (yi Yoveevsya) MR define doo(xsy) = max. bss yl ‘Then dy. is @ metric. Indeed, for z 21y 2-93) ER", we have ba ls baal +e sl max fs 3] + max [a — ih from which it follows that max [x — |= max, [ay — 3] +, mua — Remark, Examples (i), (ii) and (iv) show that we can define more than one distance function, ie. diferent metrics inthe set of ordered n-tuples. (v) Let X be any nonempty set whatsoever and let don =f ead Tt may be easily verified that dis a metric on X. Iis called the discrete metric on X. If Yis a nonempty subset of X, the metric induced on Y by dis the discrete metric on Sequence spaces provide natural extensions of Examples (i), (ii) and (iv). (vi) The space of all bounded sequences. Let X be the set of all bounded sequences of numbers, ie. all infinite sequences = (xan) = hier such that sup] <8 If x =(x};=) and y = {yi}; belong to X, we introduce the distance (xy) = sup ~ It is clear that is a metric on X. Indeed, if 2 = {z= is in X, then 1.2. Metric Spaces 31 ba - yl 0 and N >0 such that sup |f(x)| = M and sup|g(x)|=N. a re Ie follows that sup f(x) ~ g(x)| < oo. Indeed, [fd — g(x)| = f(29| + [go] = peg [fea] + sup Iso), and so 0 sup [f) ~ s(2)| = M+. Define atfig) = sup fA) — ge, fag € 1S). 32 1. Basie Concepts Evidently, d(f,g)=0,d(f.g) =0 if and only if f(x) = g(x) for all x€ S$ and aif.) = dtg,f). It remains to verify the triangle inequality for (5). By the triangle inequality for R, we have U2) ~ e629] 5 If(x) — htx)] + [AC — e629] sup [f(o) ~ hls] + sup J) ~ (0) = lf) + dhs) and so fg) = AFI) + Ag), for all fg, € (5). The metric d is called the uniform metsic (or supremum. metric) (x) The space of continuous funetions. Tet X he the set of all continnows functions defined on [4b], an interval in R. For fg € X, define AF, 8) = SBP xciai| fC) ~ 80] ‘The measure of distance between the functions fand gis the largest vertical distance between their graphs (see Figure 1.1). Since the dillerence of two continuous functions is continuous, the composition of two continuous functions is continu: ‘ous, and a continuous function defined on the closed and bounded interval [a] is bounded, it follows that dif, ¢) € R forall f,g € X. It may be verified as in Example (ili) that isa metric on X. The space X with metric « defined as ahave is denoted by Cla]. All that we have said is valid whether all complex-valued continuous functions are taken into consideration or only real-valued ones are. When it is necessary to specify which, we write Celab] or Gyla, b]. Note that Cla, b] C ‘B(a, b] and the metric described here is the one induced by the metric in Example (vili) and is also called the uniform metric (or supremum metric). (x) The st ofall continuous functions on [4,2] can also beequipped with the metric ag) fire = gla)|ae 7 TNL Figune 1 1.2. Metric Spaces 3 Figure 1.2 “The measure of stance between the functions fand ¢ represents the area berween their graphs, indicated by shading in Figure 12. If f.g€Clabl, then Lf — gl € Cla.b], and the integral defining d(f,g) is finite It may be easily verified that disa metric on Cla, b]. We note thatthe continuity of the functions enters into the verification of the “only if” part of (MS2). (ai) Let X = RU foo} U{ ~ oo}. Define f:X — R by the rule lee if 00 0, there exists @ natural number ry such that (px) < & whenever n= ry. In this case, we also say that {x} converges tox, and write it in symbols as x_ —> x. If there is no such x, we say that the sequence diverges. A sequence is said to be convergent if it converges to some limit, and divergent otherwise. Remark 1, By comparing the above with the definition of convergence in R (or C), we find that sj — if and only if lity dln) ~ 0, where d denotes the usual metric in R (or ©). Remark 2. In case there are two or more metrics on the set X, then itis necessary to specify which metrics intended to be used in applying the definition of convergence. Remark 3. If x €X and x" € X are such that x, — x and xy —> x’, then x=’. Indeed, for ¢ > 0, there exist natural numbers ry and my such that dxy,x) < 6/2 whenever m= mand d(%x’) <¢/2 whenever n = my. Consequently, (xx) 5 dlsnx) +dl%u¥) 0, there exists an integer fy(e) such that |x{® —5| Obegiven. For each j there exists 2 postive integer by such that WP sl < a tor keh, Consequently, d(x, x) (SH sr) <¢ fork ¥ 0, there exists an integer ky such that [x —ag] 0 and each j = I,-.-s7 there exists an integer such that iM —s] 0, there exists an integer ma) such that 1_ "=a asa whenever = ne), and so for =m) and for all j= 1,2,.... Since @> 0 is arbitrary, it follows that lim” — a for each j. On the other hand, let myx." — x, for each j. Consider any ¢ > 0, There exists an integer mt such thet Consequently, 1 ig —al Gd — eal ee 1b PALE al Sets 21+” al 1.3, Sequences in Metrie Spaces a [As the first sum contains only finitely many terms and lim... 3" there exits rm) suc that iw asl Dis a] foreach j whenever n= n(). Hence, d(x, x) <6 whenever n= ny(o)- Thus, convergence in the space of all sequences is coordinatewise covergence. (vii) Let X=G(p=1) and let dlxy) = x= txeleos and (eles are in fp. Let {x }yo1 sequeime Gp thal craves laa, 4 € by bey a", x) = lim (EH at) : Then fr > Othe exis pane meer) uh ha forall n= m(e). Since e > Os arbitrary, it follows thatlim. af” foreach k ‘The converse is, however, not true, asthe following example shows. (EE foe — yf)", where taf" lketlnnr de a Let xf = x4 Bin» where 0 ken ou {een Clearly, Ix" —a|= 8 =0 ifm >k “ rs ‘Consequently, limy a. x for each k. However, for all. Thus, s+ x in the fy-metric, The following does, however, hold. ats!) 2) ‘Theorem 1.3.4. Let (x!"},21 be a sequence in fy such that lim). xf" = x for each f, where x= {a,h,2 is an element of f,. Suppose also that for every > 0 there exists an integer m(e) such that (eee ‘Then limy oe d(x”, x) = 0. 2 1. Basie Concepts Proof. Consider any £ > 0, Then for m= mg(¢/4), we have 7 (= wer) 0, ithas been shown thet limy_.o. d(¥",x) =0. Examples 1.35. (i) Let X = Cla,b], the space ofall continuous functions on [a,b It does not matter whether we consider only real-valued functions or also allow complex-valued functions to be in Cla, b). Define + dif.) -| VG) = glade, A sequence {fyyz in X converges to f € X if and only if Jim dof) = fim, J IG) ~ Fld = 0 Convergence in this metric does not imply pointwise convergence, as the following ‘example shows: Let f(x) = 1 for all x € [0, 1] and nx when Osxs2 Sul) = 1 . 1 when? < xs ‘The functions f are illustrated in ig, 14. Now, Fics —sorite= [0 = mode = $0 a8 noo So dp f) = 0.5 n+ 00, Le. fy =f im the metric space. But f,(0) = 0 for all n while f(0) = 1, so that f, / f pointwise on [0,1] (0.1) (1,0) Prune LA

You might also like