SECTION 147 MAXIMUM AND MINIMUM VALUES 971 This can be confirmed geometrically from the graph of £ which isthe elliptic paraboloid with vertex (1, 3,4) shown in Figure 2. - Find the extreme values of f(x,y) = y?— SOLUTION Since &= -2xand f= 2y, the only critical point is (0, 0). Notice that for points on the x-axis we have y= 0,0 fx, ) = =x" <0 (if x 0), However, for points on the axis we have x= 0, so f(x,y) = y* > 0 (if y# 0). Thus every disk ‘with center (0,0) contains points where f takes positive values as well as points where takes negative values. Therefore (0,0) = O can't be an extreme value for f so has no extreme value, -— Example 2 illustrates the fact that a function need not have a maximum or minimum, value ata critical point. Figure 3 shows how this is possible. The graph of Fis the hyper FIGURE 3 bolic paraboloid = y* — x7, which has a horizontal tangent plane (2 ~ 0) at the origin. raytot You can sce that f(0, 0) ~ 0 isa maximum in the direction of the x-axis but a minimum in the direction of the y-axis. Near the origin the graph has the shape of a saddle and so (0, 0) is called a sade point of f ‘A mountain pass also has the shape ofa saddle. As the photograph ofthe geological for mation illustrates, for people hiking in one direction the saddle point is the lowest point on their route, while for those traveling in a different direction the saddle point isthe highest point. ‘We need to be able to determine whether or not a function has an extreme value ata crit ical point. The following test, which is proved at the end of this section, is analogous to the ‘Second Derivative Test for functions of one variable. [B] Second Derivatives Test Suppose the second partial derivatives of f are contin ‘uous on a disk with center (a, 6), and suppose that fila, 8) = Oand f(a, 6) = 0 Idhat is, (a, 6) isa critical point of f. Let Dom Da, B) = fala, B) flab) — (Sepa, BP {a) If D> Oand f(a, B) > 0, then f(a, 6) is a local minimum, (0) If D> Oand f(a, 8) < 0, then fa, 6) is a local maximum, {c) IED <0, then f(a, 6) isnot a local maximum or minimum. NOTE 1 In ease () the point (2,8) is called a saddle point of Fand the graph of f ‘erosses its tangent plane at (2, 2). NOTE 2 If D = 0, the test givesno information: could have a local maximum or local ‘minimum at (a, 6), of (a, 6) could be a saddle point of £ NOTE 3 To remember the formula for D. t's helpful to write it as a determinant far fey Nin fy faa fy ~ esl u Find the local maximum and minimum values and saddle points of Haya at = day + SOLUTION We first locate the critical points fade dy dy ae ‘Setting these partial derivatives equal to 0, we obtain the equations B-ys0 and ap 2010 Ce ering i Rn Nay mee sna inal tint De eight mei pr ney berm eink nes tw Sel tt ye may se oa heh rem sen se ne ep pores972 CHAPTER 14 PARTIAL DERIVATIVES FIGURE 4 zaxttytdyy tl ‘contour map of th featin Zin xanga 3s somninFigue 5. The lvl conas re” (1, 1) and (-1, “D) ae ulin shpe a india ‘hat as we rave aay fom (1,1) (1, =) inary rectan tha vl of Fare ies, Ths evel ves pear (0, 0), onthe heh, reremblyprils, Thay oa that a we ‘nave away om te arg where tho value of F is) ha values ff oceas in soe Grctins butinrease nha datos, Tasha como map suggests th presen of ha minina ard sade pant hat we funda ape FIGURES in Module 14:7 youcan use contour mens ‘weit the atone af erie pot, ‘To solve these equations we substitute y= »° from the first equation into the second one. Thi one ade = 1) = ate! = DUrt + = x6? = Dok + Ot + so there ae three real roots: x = 0, 1, “1. The three eritical points are (0, 0). (1, and (—1, “1 [Next we calculate the second partial derivatives and DXx, fea ltt fy 4 fy = ay? DAW = finby ~ (fe)? = Mae? — 16 Since 1{0, 0) ~ 16 <0, it follows from case (¢) of the Second Derivatives Test that the origin is a saddle point; that is, has no local maximum of minimum at (0,0). Since DU, 1) = 128 > O and fa(1, 1) = 12 > 0, we see from case (a) of the test that 1(1, 1) = 1 isa focal minimum. Similarly, we have D(—1, 1) = 128 > 0 and fa(=1, =1) = 12 > 0,50 =I, =1) = =1 is also a local minimum. ‘The graph of is shown in Figure 4 = Find and classify the critical points of the function Mag) = 10x? y = 5x? = Ay? = ‘Also find the highest point on the graph of £ SOLUTION The fest order partial derivatives are f= xy ~ Wx 4x? f= 10x? ~ By ~ 8y* So to find the critical points we need to solve the equations ow 2x(ly - 5 = 2x") =0 H 5x —4y—4y=0 From Equation 4 we see that either 0 or toy 5— 2x70 ap 2010 Ce ering i Rn Nay mee sna inal tint De eight mei pr ney berm eink nes tw Sel tt ye may se oa heh rem sen se ne ep poresSECTION 147 MAXIMUM AND MINIMUM VALUES 973, Inthe frst case (x = 0}, Equation 5 becomes ~4y(1 + 7) = 0, $0 y = O and we hhave the critical point (0,0), the second case (I0y ~ 5 ~ 2x? = 0), we get ita} w= Sy 25 ‘and, putting this in Equation 5, we have 25y ~ 12.5 ~ 4y— 4y? = 0, So we have to solve the cubic equation o Ay! — ly + 12.5, Using a graphing calculator or computer to graph the function ay) = ay’ — 2ly + 125 as in Figure 6, we see that Equation 7 has three real roots. By zooming in, we can find 3 2.7 the roots to four decimal places: yr 725452 y= 0.6468 y= 1.8984 FIGURE 6 (Alternatively, we could have used Newton’s method or a rootfinder to locate these roots) From Equation 6, che corresponding x-values are given by x= ty5y= 03 If y = ~2.5452, then xhas no corresponding real values. If y = 0.6468, then x= +0.8567, If y= 1.8984, then x= +2.6442. So we have a total of five critical points, which are analyzed in the following chart. All quantities are rounded to two ‘decimal places. Critical point | Valueot | fa D Conclusion (0,0) 0.00 = 10.00 80.00 | Tocal maximum (£2.64, 1.90) 850 3593 | 2488.72 | tocal maximum (086,06 | -148 —se7 | -18764 | saddle point Figures 7 and 8 give two views of the graph of and we see that the surface opens downward, [This can also be seen from the expression for fx, 9: The dominant terms are =x! ~ 2y! when |x| and |y/ are large.) Comparing the values of fat its local maxi- ‘mum points, we see thatthe absolute maximum value of fis /(=2.64, 1.90) ~ 8.50. In ‘other words, the highest points on the graph of F are (+2.64, 1.90, 8.50) HEE i 149 hows sera ais . cfenfaes Te xtocan goes 8 ‘is amember of one of thase families. x ’ FIGURE? FIGURE 8 — ap 2010 Ce ering i Rn Nay mee sna inal tint De eight mei pr ney berm eink nes tw Sel tt ye may se oa heh rem sen se ne ep pores(@) Closed sets —~ al f ' i Qo ro — (0 Set that are no closed FIGURE 11 ih 010 Cap Leng Alipay then cl pn wie Da crn is in pry se tw Sel tt ye may se oa heh rem sen se ne ep pores SECTION 147 MAXIM AND MINIMUM VALUES 975 Solving this equation for z, we get 2 = (12 — ay)/[2(x + 9] so the expression for V becomes yey Bow. y= ey YIRFW URFD ‘We compute the partial derivatives: av aw av _ 82 axy— Y) or yy tat If Vis a maximum, then a W/ax = aV/ay = 0, but x= 0 or y= 0 gives V= 0, so we ‘must salve the equations W-2y-e=0 12-day y= ‘These imply that x? = y# and so x= y. (Note that x and y must both be positive in this, problem) IFwe put x= yin either equation we get 12 — 3x = 0, which gives x= 2, Y= 2and 2 = (12 ~ 2* 2/22 +2) = 1 ‘We could use the Second Derivatives Test to show that this gives a local maximum ‘of Vor we could simply argue from the physical nature of this problem that there must be an absolute maximum volume, which has to occur at a critical point of F, so it must ‘oceur when x= 2, y= 2,2 = 1, Then V= 2-2-1 = 4, so the maximum volume of the box is 4 m’ — IE Absolute Maximum and Minimum Values Fora function Fof one variable, the Extreme Value Theorem says that ifs continuous on closed interval (a, 6) then has an absolute minimum value and an absolute maximum value. According to the Closed Interval Method in Section 3.1, we found these by evalu ating F not only atthe critical numbers but also a the endpoints and b ‘There i a similar situation for functions of two varlables. Just asa closed interval con tains its endpoints a closed set in Ris one that contains all its boundary points. [A bound ary point of Dis a point (a, 2) such tha every disk with center (a, 2) contains points in D and also points not in D.| For instance, the disk D=(ay|et yal) Which consists ofall points on and inside the circle x* + y* = 1, isa closed set because it contains all of its boundary points (which are the points on the circle x* + y* = 1). Butif ‘even one point on the boundary curve were omitted, the set would not be closed. (See Figure 11) ‘A bounded set in 2" is one that is contained within some disk. In other words, itis finite inextent. Then, in terms of closed and bounded ses, we can stat the following counterpart of the Extreme Value Theorem in two dimensions. [E] fxtreme Value Theorem for Functions of Two Variables If is continuous on a closed, bounded set Din R®, then f attains an absolute maximum value f(x, 92) and an absolute minimum value flv, 2) at some points (x,y) and (x 2) nD. sok ie sCot)976 CHAPTER 14 PARTIAL DERIVATIVES » 0.2] be ry Zao FIGURE 12, 9 ° L 2 ° 4, FIGURE 13 Sey) Po Day + 2y ‘To find the extreme values guaranteed by Theorem 8, we note that, by Theorem 2, if f hhas an extreme value at (x:..1), then (i,j) is either a critical point of f or a boundary point of D. Thus we have the following extension of the Closed Interval Method. [B] To find the absolute maximum and minimum values ofa continous function Ton a closed, bounded set D: 1. Find the values of fat the critical points of Fin D. 2. Find the extreme values of on the boundary of D. 3. The largest ofthe values from steps 1 and 2 is the absolute maximum value; the smallest of these values is the absolute minimum value. Find the absolute maximum and minimum values of the function fx, 9) = = 2ay + 2yon the rectangle D = (x,y) [O= x= 3.0 y= 2). SOLUTION Since fis polynomial, itis continuous on the closed, bounded rect so Theorem 8 tells us there is both an absolute maximum and an absotute min According to step 1 in [J], we frst find the critical points. These occur when le D, fe ax-n0 Ge -2yt2— so the only critical point is (1, 1), and the value of f there is /(1, 1) = 1 In step 2 we look at the values of fon the boundary of D, which consists of the four line segments L;, L2, Zs, Ly shown in Figure 12. On L, we have y = 0 and fx Q=2 0x3 ‘This isan increasing function of x; so its minimum value is (0, 0) ~ O and its maxi- ‘mum value is (3,0) = 9. On La we have x= 3 andl $B) -9-4y OS y=? ‘This is a decreasing function of y, so its maximum value is £13, 0) = 9 and its minimum value is (3, 2) = 1. On Ly we have y= 2 and Ax Q=v— deed OS eS3 By the methods of Chapter 3, or simply by observing that f(x, 2) = (x — 2), we see that the minimum value of this function is f(2, 2) = 0 and the maximum value is (0, 2) = 4. Finally, on La we have x= 0 and MOy— ty OSys2 ‘with maximum value 0,2) = 4 and minimum value £0, 0) = 0. Thus, on the bound- ary, the minimum value of fis 0 and the maximum is 9. In step 3 we compare these values with the value f(1, 1) = 1 at the critical point and conclude that the absolute maximum value of fon Dis 3,0) = 9 and the absolute ‘minimum value is £(0, 0) = f(2, 2) = 0. Figure 13 shows the graph of £ - ap 2010 Ce ering i Rn Nay mee sna inal tint De eight mei pr ney berm eink nes tw Sel tt ye may se oa heh rem sen se ne ep pores