Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
78 views64 pages

Referensi 1 - Purcell

Uploaded by

nawra illona
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
78 views64 pages

Referensi 1 - Purcell

Uploaded by

nawra illona
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 64
CHAPTER 1 3.2 Maxima and Minima Monotonicity and Concavity Local Extrema and Extrema on Open Intervals Practical Problems. Graphing Functions Using Calculus ‘The Mean Value Theorem for Derivatives Solving Equations Numerically Antiderivatives Introduction to Differential Equations (0 0,:2),m0 max oe mia On 1, 3}, max = 1, nin = (0n(1,3) pom, is Figure 2 Applications of the Derivative 3.1 Maxima and Minima Often in fife, we are faced with the problem of finding the best way to do some- thing, For example, a farmer wants to choose the mix of crops that is likely to pro- duce the largest profit. A doctor wishes to select the smallest dosage of a drug that, will cure a certain disease. A manufacturer would like to minimize the cost of di Uributing its products, Often such a problem can be formulated so that it involy ‘maximizing or minimizing a function over a specified set. If $0, the methods of cal- culus provide a powerful tool for solving the problem. ‘Suppose then that we are given a function f(x) and a domain $ as in Figure 1 ‘We now pose three questions: 1. Does f(x) have a maximum or minimum value on $? 2. If it does have a maximum or a minimum, where are they attained? 3. I they exist, what are the maximum and minimum values? Answering these questions is the principal goal of this section, We begin by intro- ducing a precise vocabulary: Definition LetS, the domain of {contain the point c. We say that (@) PenstiemaximumvaRTE of fost Gi) -F(e)istheminimumvalue of fon S forall in il) eE—_—_— if itis either the maximum value or the (iv)_ the function we want to maximize or'minimize is the objective function, ‘The Existence Question Does fhave a maximum (or minimum) value on 8? ‘The answer depends first of all on the set S. Consider f(x) = 1/xon S = (0, 0): it has neither a maximum value nor a minimum value (Figure 2). On the other hhand, the same funetion on § = [1, 3] has the maximum value of f(1) = Land the Lons minimum value of f(3) (1,3), f has mo maximum value and the min- imum value of (3, ‘The answer also depends on the type of function. Consider the discontinuous function g (Figure 3) defined by fz ifl=x<2 v-2 if2 ¢, then Figure 7 Terminology Notice the way that tems are used in Example 3.The maximum is 1, whieh is equal to f(—!) and fC). We say that the maximum is attained at! and at 1, Similarly the min rum is 4, which i attained at 2 Soya 4 doo Figure 9 Section 3.1 Maxima and Minima 153 @ $2 -10 og But f'(c) exists because ¢ is not a singular point. Consequently, when we let xin (1) and xc" in (2), we obtain, respectively, f"(c) = Oand f"(e) = 0. ‘We conclude that f”(c) = 0, as desired. ‘The case where f(c) is the minimum value is handled similacly. . In the proof just given, we used the fact that the inequality = is preserved under the operation of taking limits, What Are the Extreme Values? _ In view of Theorems A and B, we can now Bef seas 9 eee Toh ecstatic to Df acoqnianaa fol clad lore Step I: Find the ertcal points of fon 1 Sf Tvs 7 acho teas cried pone Te Tampa thas alacant ‘povecum wines barman la theamoisaretiig [EEXAMPLE2] Find the maximum and minimum values of f(x) = 22 on ra 2 - 1-2.2} A= adeboed yy (3)-2) aenl gots , SOLUTION The derivative is f'(x) = 3x2, which is defined on (2,2) and is Ee ren esate lminieess teastone 1 ent n-ne pal x = =2and x ~ 2, Evaluating fat the critical points yields f(-2) = -8, f(0) = 0, and (2) = 8. Thus, the maximum value of f is 8 (attained at x = 2) and the minimum is ~8 (attained atx = ~2), . Notice that in Example 2, "(0) = 0, but fdid not attain a minimum or a max- imum at x = 0. This does not contradict Theorem B. Theorem B does not say that ite eral point then (6) isa minimum of maximo i says thaihei-a Fin the maximum and miniznam values of f(a) = 20 +32 SOLUTION In Example 1, we identified ~}, 0, 1, and 2 as the evtical points. Now f(—3) = 1. f(0) = 0, F) ind f(2) = —4. Thus, the maximum value is 1 (attained at both x = —Land x = 1),and the minimum value is ~4 (attained at 2).The graph of fis shown in Figure 7. . ‘The function F(x) = 2° is continuous everywhere, Find its ‘maximum and minimum values on [=1, 2]. SOLUTION F'(x “3, which is never 0. However, F"(0) does not exist, so 0 is critica point, as are the end points ~1 and2, Now F(=1) = 1, F(0) = 0.and F(2) = Wa ~ 1.59. Thus, the maximum value is ¥/4: the minimum value is 0. The graph is shown in Figure 8. . Find the maximum and minimum values of f(1) =x + 2cos.x on[=#, 22. SOLUTION Figure 9 shows a plot of y= f(x). The derivative is $'(x) = 1 = 2sin.x, which is defined on (—x, 27) and is zero when sin x ‘The only values of x in the interval [—7, 27] that satisfy sin x = 1/2 are x and x = 5n/6, These two numbers, together with the end points —7 and 2, are the crtieal points. Now, evaluate fat each critical point: 184 Chapter 3 Applications of the Derivative A- - 9s S14 Sajo) = Va +E = 226 J(Sa/6) =-V3-+7Z 089 f(2e) = 2+ 2 = 828 ‘Thus, -2 — {attained at x = 27), Concepts Review LA both a maxi 2. The term ‘minimum value, function on a interval will always have jum value anda minimum value on that interval value denotes either a maximum or a is the minimum (attained at x = ~z),and the maximum is 2 + 2 3. A function can attain an extreme value only ata critical ‘point. Critical points are of three types and. 4. A stationary point for fis.a number e such that ‘singular point for fis & number c such that Problem Set 3.1 In Problems 1-4, fina alleritical points and ind the minimum and ‘maximum of the function. Each function has domain [2,4] hy 2 ? 11 10) = snot = [-2.2| 1, 99) = sine coves = (0471 1B. a(s) = he 1 20, a. ST = [1,32] rt FPP 2 2B. 3 wv = fi 24. } re % sa Taos ae 24 1 3 aF 26. le) = In Problems 5-26, identify the crtial points and find the maxi ‘mum value and minimum value on the given interval fl) = 2 + 4 +441 = [4,0] 6 hy tad 2 8 Gls) = Hae + ae LaF = (3.3) 9. fl) = 28 = au + 1:1 = (4) Mine Sketch the graph 1, f(x) = —3e+ r= [4 mh ry == 1.3) 2 6) =a =a 1B fly) = xt 2 T= [2.2] py Be 420 k= 3.2] 1 T+ 15. g(x) (09, &) Hints Sketch the graph, 16 f(x) = Teen I= (1.8) AN, tenis the eritical points and find the extreme values on the interval [—1, 5] for each function: (a) fla) = 8-6 +2) glx) = [FGI ‘Zs, tdemtiy the critical points and find the extreme values on the interval [-1, 5] for each function: (9) F(a) = cos + xsin +2 (b) gx) = [FO] In Problems 20-36, sketch the graph of a function with the given properties. 20, fis differen (0F6 (attained when x = 0), Additionally 30. fis differentiable, has domain [0,6], reaches a maximum ‘of (attained when x = 6) and a minimum of ~2 (attained when x= 1).Aduitionally,x = 2,3, 4,5 ase stationary points, 31. f is continuous, but not necessarily differentiable, has domain [0, 6], reaches a maximum of 6 (attained when x = 5), and a minimum of 2 (attained when x = 3). Additionally, x = 1 and x = Sare the only stationary points ble, has domain [0,6], reaches a maximum ) and a minimum of O (attained when isa stationary point 32. f is continuous, but not necessarily 0.’Then it seems reasonable that the abject will con- tinue 10 move to the rigl ther words, jese observations do not prove Theorem A, nut they make the result plausible, We postpone a rigorous proot until Section 36 Monotonicity Theorem differentiable at every interior point of 1469S" alt x interior to J then fisinereasingOn all interior to Zhen fis deereusing on This theorem usually allows us to determine precisely where a differentiable function increases and where it decreases. It is a matter of solving two inequalities. EXAMPLE 1] If f(x) = 2x°- and where itis decreasing, SOLUTION We begin by finding the derivative of f. f'(x) = 6x2 ~ 6x = 12 = 6(x + (x - 2) 0 renro } rereo L 0 Figure 2 x? = 12x +7, find where f is increasing 156 Chapter 3 Applications of the Derivative Valucso f° Figure 3 fis) = 28-38 12047 Figure 4 Values af yp Figure 5 Increasing bt wig Figure 6 We need to determine where (e+e -2)>0 and also where (e#1)e-2) <0 This problem was discussed in great detail in Section 0.2, section worth reviewing now. The split points are —I and 2; they split the a-axis into three intervals (-ce, =1), (1,2), and (2, 00). Using the test points ~2,0, and 3, we conclude that f"(x) > 0 on the frst and last of these intervals and that f"(x) <0 on the middle interval (Figure 3). Thus, by Theorem A, fis increasing on (—20, ~1] and [2 co}; itis decreasing on [—1, 2], Note that the theorem allows us to include the end points ofthese intervals,even though f"(x) = Oat those points The graph of f is shown in Figure 4 . EXAMPLE 2] Determine where g(x) = x/(1 + x°) is increasing and where it is decreasing. SOLUTION _(=x0+x) sie (+e +eyF Since the denominator is always positive, ”(x) has the same sign as the numerator (1 x)(1 + x), The split points, =I and 1, determine the three intervals (00, “1), (=, 1), and (1, 26). When we test them, we find that g'(x) < On the first and last of these intervals and that g’(x) > 0 on the middle one (Figure 5). We conclude trom Theorem A that g is decreasing on (~00, —1] and [1,c0) and that itis increasing on [—1, 1]. We postpone graphing g until later, but if you want to see the graph, urn to Figure II and Example 4 . ‘The Second Derivative and Concavity A function may be inereasing and still have a very wiggly graph (Figure 6). To analyze wiggles, we need to study hhow the tangent line turns as we move from left to right along the graph. Ifthe tan- gent line turns steadily in the counterclockwise direction, we say that the graph is concave up; if the tangent turns in the clockwise direction, the graph is concave down. Both definitions are better stated in terms of funetions and their derivatives. Definition Let fbe differentiable on an open interval I. We say that f(as well as its graph) is concave up on Ff f" is increasing on /, and we say that fis concave down on I if f" is decreasing on L ‘The diagrams in Figure 7 will help to clarify these notions. Note that a curve that is concave up is shaped like a cup. YF ineresing: Concave pf Soewasing: Concave down Concave wp Coneve down Figure 7 Conditions in Theorems A and B The conditions regarding the de tives in Theorems A and B are suff ‘cient t guarantee the conclusions stated, These conditions are not, however, necessary Its possible that a function is increasing on some interval even though the derivative isn't always positive on that inter val. we consider the function fx) = x over the interval [~4, 4] ‘We note that its increasing but is derivative is not always positive on that interval (f"(0) = 0). The func- tion g(x) = x*iscancave up on the interval [4,4], but the second de- 12x", is mot always foro oe th = 2 r o + Figure 8 Section 3.2 Monotonicity and Concavity 187 In view of Theorem A, we have a simple criterion for deciding where a curve is concave up and where itis concave down, We simp! derivative of fi decreasing if, keep in mind that the second the first derivative of f'. Thus, f” is increasing if f” is positive: itis is negative, Concavity Theorem Let fhe twice differentiable on the open interval /. () IE F*(x) > O for all.vin J, then fis concave up on 1 (i) If f"(x) < O for all.xin Z,then fis concave down on I For most functions, this theorem reduces the problem of determining concav- ity to the problem of solving inequalities. By now we are experts at this, EXAMPLES) Where is f(x) = {1° - x? - 3x + 4 increasing, decreasing, ‘concave up, and concave down? SOLUTION P(x) =P =e 3 = (x + IY -3) I(x) = 2x 2=2x-1) By solving the inequalities (x +1)(y-3)>0 and its opposite, (x + 1)lr = 3) < 0, we conclude that / is increasing on (—20, ~1] and [3, 00) and decreasing on [1,3] (Figure 8). Similarly, solving 2(x — 1) > 0 and 2(x — 1) <0 shows that f is concave up on (1,00) and concave down on (—00, 1). The graph of fis shown in Figure 9, . EXAMPLE 4] Where is g(x) = x/(1 + x°) concave up and where is it con- cave down? Sketch the graph of g. SOLUTION We began our study of this function in Example 2, There we earned that g is decreasing on (—o°, ~1] and [1, 0°) and increasing on [=1, 1}. To analyze concavity, we calculate 3", = (= YQ) + F)2) (+x) _ + NG + P28) = = EY) at+ey 2x} = 6x a+eyp Since the denominator is always positive, we need only solve x(x? — 3) > 0 and its opposite, Te spit points are ~V3,0, and V3, These thre split points deter mine four intervals. After testing them (Figure 10) we conclude that gis concave up on (~V3.0) and (V3, c0) and that it is concave dawn on (~c0, V3) and (v3). 158 Chapter 3 Applications of the Derivative a in Figure 12 Figure 13, To sketch the graph of g, we make use of all the information obtained so far, plus the fact that g is an odd function whose graph is symmetric with respect to the origin (Figure 11). = teeveasing ——— >] increasing fe deereasing ——e concave down al _ernsne ee of conc 3p we ° ve Figure 11 Suppose that water is poured into the conical container, as shown in Figure 12, at the constant rate of | cubie inch per second, Determine the height ofthe water asa function of time cand plot (from time r= O until the time that te container s fl SOLUTION ) Before we solve this problem, fet’ think about what the graph will look like. At firs. the height wil increase rapidly. since it takes very little water to fil the bottom ofthe cone. As the container fill up, the height will increase less rapidly. What do these statements say about the function h(0), its derivative h'(1), and its second derivative (1)? Since the water is steadily pouring in, the height will always increase, so 1"() will be postive. The height will increase mote slowly 4s the water level rises. Thus the funtion h(t) is deereasing so h(t) is negative. The graph of h() is therefore increasing (because h(t) is postive) and concave down (because h"(t) is negative). [Now once we have an intuitive idea about what the graph should look like (in- creasing and concave down), et’ solve the problem analytically. The volume of a right circular cone is V = Lr%h, where V. rand fare all funetions of time. The functions hand rare related: notice the similar triangles in Figuce 13, Using prop erties of similar triangles, we have 4 Thus,r = h/4. The volume of the water inside the cone is thus On the other band, since water is flowing into the container atthe rate of} cubie inch per second, the volume at time ris V = 41, where r is measured in seconds. Equating these two expressions for V gives, nes sxpweuven= Ye =r eonsy mii rman enone A ecbce st einen onraceR acl be ane la MRO Figure 15 Figure 16 Terminology While a function's minimum oF ‘maximum is a number, an inflestion point is always an ordered pair I) y falas Figure 18 Section 3.2 Monotonicity and Coneavity 189 ‘The first and second derivatives of h are Wi) =D. which is positive, and (= De which is negative, The graph of f() is shown in Figure 14. As expected, the graph. of h is increasing and concave down, . AA news ageney reported in May 2008 that unemployment in eastern Asia was continuing to increase at an increasing ate. On the other hand, the price of food was increasing, bt ata slower rate than before Interpret these ftatementain (enne of increasing) decreasing functions and concavity. SOLUTION Let u = f(s) denote the number of people unemployed at time 1. Although « actually jumps by unit amounts, we will follow standard practice in representing w by a smooth curve as in Figure 15. To say unemployment is, increasing is to say that duu/dr > 0."To say that itis increasing at an increasing rate is to say that the function du/dt is increasing: but this means that the derivative of du/dt must be positive. Thus, @°u/dr? > 0. In Figure 15, notice that the slope of the tangent line increases as r increases. Unemployment is increasing and concave up. ‘Similarly, if p = g(1) represents the price of food (eg., the typical cost of one day's groceries for one person) at time 4, then dp/dt is positive but decreasing. ‘Thus, the derivative of dp/dt is negative, so dpjdt’ <0. In Figure 16, notice that, the slope of the tangent line decreases as increases. The price of food is inereasing but concave down, . Inflection Points Let/be continuous at e.We call (c, f(c)) an inflection poi of the graph of fit fis concave up on one side of e and concave down on the other side. The graph in Figure 17 indicates a number of possibilities. Intestion Innestion aa rs | Conse Canave : chm | Concave ‘tow Concave i Concave oD Figure 17 {As you might guess,poinis where f*(x) = 0 or where {"(x) does not exist are te candidates for points of infection. We use the word candidate deliberately. ust 5a candidate for politcal office may fil to be elected, so. for example, may @ point where f"(x) = 0 fail to be a point of inflection. Consider f(x) “which fas the graph shown in Figure 18 lis true that f"(0) ~ 0: yet the origin not point of inflection, Therefore. in searching fr infection points we begin by ident- Fying those points where f(x) = 0 (and where f*(x) does not exis) Then e check to see it they really ae inflection points Look back atthe graph in Example 4, You will se that it has three inflection points They are (=\V3, ~V3/4). (0.0),and (V3, V3/4) 160 Chapter 3 Applications of the Derivative SOLUTION ‘The second derivative, F(x), is never 0; however, it point (0, 2) is an inflection point since F"(x) > 0 for x <0 and F"(x) <0 for x > 0, The graph is sketched in Figure 19. . EXAMPLE 7] Find all points of inflection of F(x) = x" + 2. 1 3x 2 POs a9 POS Concepts Review 1. If F'(x) > everywhere, then fis __ everywhere; if #708) > Ceverywhere, then fis ___ everywhere. 2 If __ and _on an open interval J, then fis both increasing and eoncave down on 1 3. A point on the graph of a continuous function where the ‘concavity changes is called 4. In trying to locate the inflection points for the graph of a function J we should look at numbers ¢, where either —__ or Problem Set 3.2 In Problems 1-10, use the Monotonicity Theorem to find where the ven function is inereasng and where its decreasing 1 09 2 ay (xt He BwQ=P+4-3 A flee 5 Gx) Dx 6 f@) 2 ame 2 8 0) © 9. HW) = sin. 56-5 2e 10, R(0) = cos 6,0 = 0 = 2a In Problems I-18, use the Concavity Theorem to determine Where the given function is concave up and where itis concave down. Also find all inflection points: AL fla) = (x=? 2, Giw) 13, Ty) = 38 = 184 18, g(x) = x4 - 6x4 — 4a 16, f(a) = 38 + 8 18 G(x) = 2 + 12sine In Problems 19-28, determine where the graph of the given fune tion is increasing, decreasing, concave up, and concave down. Then sketch the graph (see Example 4). 1, fx) == 12x +1 0, glx) = 4x3 — 3x2 — 6x + 12 BL glx) = 3x6 4x 222 PU 23, G(x) = 3S Sx 2, HEX) 7 28, f(x) = Vain won [0.7] 26. g(x) = xT fx) = OMB. gs) = Be + In Problems 20-34, sketch the graph of a continuous function fom (0,61 char saisies al the stated conditions 2, (0) = 1:/(6) = 3 increasing and concave down on (0.6) Mo. (0) = 8:16 decreasing on (0, 6); inflection point at the ordered pair (2,3), concave upon (2,6) BL /(0) = 34/08) = 0/16) = 4 L(2) < Bon (0.3):/"() > Gon (3,6): F(a) > Don (0.5): f"(2) < Don (5,6) 32 f(0) = 5:F(2) = 2 £16) = 0% P(x) <00n(0.2)U2.6):F £"(x) < Bon (0,1) U(2,6):f"(x) > Bon (1,2) BB. f(0) = f(A) = 4) = 2:f(6) = 0: #1) > Don (0.2): /'() < Gon 2,4) U4, 6); PQ) = f(A) = 0:F"() > Don (G,1)UBA: (3) < on (1,3) U(4.6) H f(0) = f13) = 3:42) = 4: F(8) = 25116) = 0: #1) > Don (0.2): /'(4) < Gon (2,4) U4, 5}; PQ) = £4) = 0:/'(4) = —10n (5,6); $"(x) < Don (0,3) UE, 5):/"(x) > Bon (3,4) 235. Prove that a quadatic function has no point of inflection 36. Prove that a cubic function has exactly one point of inflection. 37, Prove that if (x) exists und is continuous on an interval Land if f"(s) #0 at all interior points of J, then either fis increasing throughout For decreasing throughout J Hint-Use the Intermediate Value Theorem to show that there cannot be two points x, and x2 of J where f' has opposite signs. 3B Suppose that fis a function whose derivative is PCa) = (8 = x + D/C + 1). Use Problem 37 to prove that f is increasing everywhere, 39, Use the Monotonicity Theorem to prove each statement o 0 and g'(x) > 0 for all x. What simple addtional conditions (if any) are needed to guarantee that: {a) f(x) + a(x) isinereasing forall {b) #€2)+@(2) isincreasing forall x; {©} fle() isinereasing forall x? 44. Suppose that f"(x) > 0 and g°(x) > 0 for all x What simple additional conditions (if any) are needed to guarantee that (a) F(2) + g(x) isconcave up forall x: (©) F(2)+4(4) is concave up forall x; (©) F(g(2)) is eoncave up forall x? Use a graphing calculator or a computer to do Problems 5-8, 48, Let (x) = sin x + c0s(x/2) on the interval / (a) Draw the g1aph of fon {b) Use this graph to estimate where f"(x) ) containing ¢ such that f(c) isthe minimum value of fon (a,b) 1S: (ii) /(©) isa local extreme value of fit itis either a local maximum or a local ‘minimum value, Section 3.3 Local Extrema and Extrema on Open Intervals 163 Where Do Local Extreme Values Occur? The Critical Point Theorem (Theorem 3.1B) holds with the phrase extreme value replaced by focal extreme value; the proof is essentially the same. Thus, the critical points (end points, sta- tionary points, and singular points) are the candidates for points where local ex- ‘tema may occur. We say candidates because we are not claiming that there must be a local extremum at every critical point, The left graph in Figure 3 makes this, > Siope Figure 3 faye [No al extreme valve Local maxium ‘Local minim ore clear, However, if the derivative is positive on one side of the critical point and negative on the other (and if the funetion is continuous), then we have a local ex- ‘tremum, as shown in the middle and right graphs of Figure 3. First Derivative Test Let fbe continuous on an open interval (a,b) that contains a critical point c (@__UEf'(x) > 0 forall xin (a,c) and f"(x) < 0 for allxin (¢,b), then f(e) is local maximum value off Gi) IEf'(x) < 0 forall x in (a,c) and f"(x) > 0 for allxin (¢,b), then f(e) is local minimum value off (ii) IF f'(x) has the same sign on both sides of c, then f(c) is not a local extreme value off. Proof of (i) Since f’(x) > 0 for all x in (a, ¢) fis increasing on (a, ¢] by the Monotonicity Theorem. Again, since f"(x) < 0 for allx in (« b),fis decreasing on fc. ). Thus, f(x) < fc) for all xin (a,b), except of course at x = ¢, We conclude that f(¢) is a local maximum, ‘The proots of (i) and (i are similar, . EXAMPLE 1] Find the local extreme values of the function f(x) = = 6x + Son (—00, 00), SOLUTION The polynomial function f is continuous everywhere, and its derivative, f’(x) = 2x = 6, exists for all x. Thus, the only critical point for fis the single solution of f’(x) = 0; that is,x = 3. Since f"(x) = 2(x ~ 3) <0 for x <3, fis decreasing on (—c0, 3]; and be- cause 2(x ~ 3) > 0 for x > 3, fis increasing on [3, cc). Therefore, by the First Derivative Test, f(3) = ~4 is a local minimum value of f. Since 3 is the only criti- cal point, there are no other extreme values. The graph of fis shown in Fi Note that f(3) is actually the (global) minimum value in this case. Find the local extreme values of f(x) =x? =a? = 3x44 ‘on (=00, 00), 164 Chapter 3 Applications of the Derivative Figure 5 Figure F00)=(sina SOLUTION Since f'(x) = x7 = 2x = 3 = (x + 1)(x = 3), the only critical points of fare —1 and 3, When we use the test points ~2, 0, and 4, we learn that, (x # 1) = 3) > 0 om (=00,=1) and (3,00) and (x + 1)(x = 3) <0 on (-1,3). By the First Derivative Test, we conclude that f(=1) = is a local ‘maximum value and that /(3) = ~S is a local minimum value (Figure 5), . Find the loeal extreme values of f(x) = (sina)! on (-#/6,22/3). SOLUTION 2eos x eet #0 3¢sin x)" fx = ‘The points 0 and 7/2 are critical points, since f"(0) does not exist and f'(/2) = 0. Now f'(x) < 0.0n (—n/6,0) and on (17/2, 2/3), while f'(x) > 0 (on (0, 7/2). By the First Derivative Test, we conclude that f(0) = 0 isa local min- imum value and that f(m/2) = 1 is a local maximum value. The graph of f is, shown in Figure 6, . ‘The Second Derivative Test There is another test for local maxima and ‘minima that is sometimes easier to apply than the First Derivative Test. I involves, evaluating the second derivative at the stationary points. It daes not apply to sin- gular points. Second Derivative Test Let f" and * exist at every point in an open interval (a, b) containing c, and suppose that f"(c) = 0. i) IF*(e) <0, then F(c) isa local maximum value off (ii) IEF*(e) > 0, then fc) isa local minimum value off Proof of (4) Its tempting to say that, since *(c) <0, fis concave downward near ¢ and to elaim that this proves (i), However, to be sure that fis concave down- ‘ward in a neighborhood of ¢, we need f"(x) < 0 in that neighborhood (not just at ©), and nothing in our hypothesis guarantees that. We must be a hit more careful By detinition and hypothesis, £89 ree But this inequality implies that f"(x) > 0 for a 0. Therefore, by the Second Derivative Test, (3) isa local minimum value. . Figure 7 Figuee 9 py sivas Section 3.3 Local Extrema and Extrema on Open Intervals 165 For f(x) = ‘Test to identify local extrema, ~ 3x +4, use the Second Derivative SOLUTION Thisis the function of Example 2. P(x) = 8 = 2-3 = (xt -3) F(x) = 2x -2 The critical points are ~1 and 3 (f"(=1) = f"(3) = 0). Since f*(—1) = ~4 and #"@) = 4,we conclude by the Second Derivative Test that f(—I) is a local mat ‘mum value and that (3) isa local minimum value, 7 Unfortunately, the Second Derivative Test sometimes fails since f(x) may be 0 at a stationary point. For both f(x) = x° and f(x) = 2%, f"(0) =0 and £7(0) = 0 (see Figure 7). The first does not have a local maximum or minimum, vvalue at 0;the second has a local minimum there. This shows that if f"(x) = Oat a stationary point we are unable to draw a conclusion about maxima or minima ‘without more information. Extrema on Open Intervals The problems that we studied in this section and in Section 3.1 often assumed that the set on which we wanted to maximize or ‘minimize a function wasa closed interval. However, the intervals that arise in prac- tice are not always closed; they are sometimes open, or even open on one end and closed on the other. We can still handle these problems if we correctly apply the theory developed in this section, Keep in mind that maximum (minimum) with no qualifying adjective means global maximum (minimum). EXAMPLE 6] Find (if any exist) the minimum and maximum values of f(x) = x4 = 4x on (~20, 00). SOLUTION PQ) = 40-4 = 400 = 1) = 4e- G2 +441) Since x? + x + 1 = 0 has no real solutions (quadratic formula), there is only one critical point, x = 1. Forx < 1, f'(x) <0, whereas for x > 1, f"(x) > 0, We con- clude that f(L) = —3 is a local minimum value for f; and since fis decreasing on the left of | and inereasing on the right of 1, it must actually be the minimum value of f. ‘The facts stated above imply that feannot have a maximum value. The graph cof Fis shown in Figure 8. 7 EXAMPLE 7] Find (if any exist) the maximum and minimum values of Gp) w= p) ‘on (0, 1). SOLUTION 2p ~ PUL — py 10) = Flea a ‘The only critical point is p = 1/2. For every value of pin the interval (0,1) the de- nominator is positive; thus, the numerator determines the sign. Ifp isin te inter- val (0, 1/2) then the numerator is negative: hence, G'(p) < 0. Similarly. if pis in the interval (1/2, 1),G'(p) > 0. Thus, by the First Derivative Test, G(1/2) ~ 4 isa local minimum. Sinee there are no end points or singular points to cheek, G(1/2) is a global minimum. There is no maximum. The graph of y = G(p) is shown in Figure 9. 166 Chapter 3 Applications of the Derivative Concepts Review 1. I fiscontinuows at, f(x) > Ona toc onthe let and Fu) < Omar oc onthe right then fle) Balad value fore 2 Af f'(x) = (x + 2)(x — 1), then f(—2) is a local safue fr fand J) sales! value fr 3. If f'(c) = 0 and f"(c) <0, we expect to find a local value for fate. 4. If f(x) = then f(0) is neither a___ nor a_, ‘even though /"(0) Problem Set 3.3 {In Problems 1-10, identify the eritical points. Then use (a) the First Derivative Test and (if possible) (6) the Second Derivative Test 10 decide which of the critical points give a local maximum and which give a local minimum. 1 fly 6x8 +4 BR f(xy= x -et ee B f(0) = sin20.0-<0-< = f(a) = fo sing 0 << Oe 4 5. VO) 6 1 sin? 0/2 <0< 0/2 find the critical points and use the test of your choice to decide which critical points give a local maximum value and which give a local minimum value. What are these local maxi- ‘mu and minimum values? ML fx) = 2 — 3x 1 g(x) = xt + x 1 M(x) = x4 208 14. fl) = (2-338 1 gi) = 9-0 16, (5) = 35 + 978 wfger-heeo a fixy= 19. A(0) = 7 oe 0<0<2n 20, g(0) = |sindl,0< 0 < 2m In Problems 21-30, find if possible, the (global) maximum and ‘minimum values of the giver function on the indicated interval BL f(x) = sin? 2x 0n [0,2] 2. fx) = jon0,20) 2B. a(x) 0" 10,26) 6 ito) Leonia.) 28, F(x) = OV ~ 4x on [0.4] 26, F(x) = Vx ~ Axon 0, ©) ‘on (0, 7/2) to on (8,00 tee) 29, H(x) = |x? = Hon[-2,2] 30, AO) = sin on [0,2] {In Problems 31-36, the fist derivative {" ix given. Find all values ‘efx that make the function f(a) a local minimum and (b) a tocat BL (y= aP 32. f(x) = (x = Ila — 2Y(x = 3) 4) 33. f(x) = (8 Or = 2)" ~ 3) 4) BL P(x) = (0 = 1 = 2 = 3 =P 38 f(x) = (x= AV = BYLA eB 6. f(x) = ax — AMX BOK AB Jn Problems 37-12, sketch a graph of a fnetion with the given properties If itis impossible to graph sucha function, hen indicate ‘this and justify your ansver 37. Fis differentiable, has domain [0, 6], and has two local ‘maxima and two local minima on (0,6) 138. is differentiable, has domain [0 6] and has three local ‘maxima and two local minima on (0,6) 39. Js continuous but not necessarily differentiable, has do- main [0,6], and has one focal minimum and one loeat maximum 00 (0.6). 40, F's continuous, but not necessarily differentiable, has do- ‘main (0,6) and has one local minimum and no local maximum on (0.6) AL. Fhas domain [0,6], ut is not necessarily continuous, and has three local maxima and no local minimum on (0,6). 42, f has domain [0,6], but is not necessarily continuous, and has two local maxima and no local minimum on (0,6). 43, Consider /(x) = Axe + Bx +C with A > 0, Show that f(x) = O forall xi and only it B° ~ 44C = 0 44, Consider f(x) = Ax’ + BY + Cx + D with A> 0. Show that f has one local maximum and one local minimum if and only if B® ~ 3AC > 0, 48. What conclusions can you draw about from the informa tion that f'(c) = f*(e) = and f*(c) > 0? Answers to Concepts Review: 1. maximum 2 maximum; minimum 3, maximum 4 local maximum: Toeal minimum: 0 3.4 Practical Problems Figuee 3 Section 3.4 Practical Problems 167 Based on the examples and the theory developed in the first three sections of this chapter, we suggest the following step-by-step method that can be applied to many practical optimization problems. Do not follow it slavishly: common sense may sometimes suggest an alternative approach or omission of some steps. Step I: Draw a picture for the problem and assign appropriate variables to the im- portant quantities. Step 2: Write a formula for the objective function @ to be maximized or mini mized in terms of the variables from step 1 Step 3: Use the conditions of the problem to eliminate all but one of these vari ables, and thereby express Q as a function of a single variable. ‘Step 4: Find the critical points (end points, stationary points, singular points). Step 5: Bither substitute the critical values into the objective function or use the theory from the last section (ie. the First and Second Derivative Tests) to deter- mine the maximum or minimum, ‘Throughout, use your intuition to get some idea of what the solution of the problem should be. For many physical problems you ean get a “ballpark” estimate Of the optimal value before you begin to carry out the details. A rectangular box is to be made from a piece of cardboard 24 inches long and 9 inches wide by cutting out identical squares from the four cor- ners and turning up the sides, as in Figure 1. Find the dimensions of the box of max- imum volume. What is this volume’? SOLUTION Let x be the width of the square to be cut out and V the volume of the resulting box. Then V = x(9 — 2x)(24 - 28 Now x cannot he less than 0 or more than 4.5. Thus, our problem is (o maximize V on [0,45]. The stationary points are found by setting dV /dx equal to O and solving the resulting equation: av dx 26x — 66x7 + 4x? 216 — 132 + 12x? = 12(18 - 1x + x4) = 129 - 2-2) =0 This gives x = 2 or x = 9, but 9 is not in the interval [0, 4.5], We see that there are only three critical points, 0, 2, and 4.5. At the end points O and 4.5, V = 0; at 2, V = 200, We conclude that the box has a maximum volume of 200 cubic inches it x = 2, that is, if the box is 20 inches long, 5 inches wide, and 2 inches deep. Itis often helpful to plot the objective function. Plotting functions can be done easily with a graphing calculator or a CAS. Figure 2 shows a plot of the function, V(x) = 216x ~ 6x7 + 4x°. When x = 0, V(x) is equal to zero. In the context of folding the box, this means that when the width of the cut-out corner is zero there is nothing to fold up, so the volume is zero. Also, when x = 4.5, the cardboard gets folded in half, so there is no base to the box; this box will also have zero volume. Thus, V(0) = 0 and V(4.5) = 0. The greatest volume must be attained for some value of x between 0 and 4.5. The graph suggests that the maximum volume occurs, when xis about 2; by using calculus, we can determine that the exact value of x that maximizes the volume of the box is x A farmer has 100 meters of wite fence with which he plan 10 build two identical adjacent pens as shown in Figure 3. What ae the dimensions of the enclosure that has maximum area? 168 Chapter 3 Applications of the Derivative Figure 4 Algebra and Geometry Whenever possible, try to view a problem from both a geometric and fn algebra point of view, Exam: ple 3 @ good example for which this kind of thinking lends insight into the proble SOLUTION Let x be the width and y the length of the total enclosure, both in, meters, Because there are 100 meters of fence, 3x + 2y = 100; that is, ‘The total area A is given by A= xy=S0x— 3x? Since there must be three sides of length x, we see that 0 = x = “®. Thus, our problem isto maximize A on [0.22], Now aa Bas0-3 dx * When we set 50 — 3 as a stationary point ‘Thus, there are thre critical points: 0, and "The two end points O and “give A= 0, while x yields A~ 41667, The desired dimensions are 1 equal to 0 and solve, we get x 16467 meters and y = 50 ~ $2) = 25 meters. | Is this answer sensible? Yes. We should expect to use more of the given fence in the y-direction than the «-direction because the former is fenced only twice, whereds the later is Fenced three times. . EXAMPLE 3] Find the dimensions of the right circular cylinder of greatest volume that can be inscribed in a given right circular cone, SOLUTION Leta be the altitude and b the radius of the base of the given cone (both constants). Denote by #, r and V the altitude, radius, and volume, respectively, of an inscribed cylinder (see Figure 4). ) Before proceeding, let’s apply some intuition. If the cylinder’s radius is close to the radius of the cone’s base, then the eylinder’s volume would be close to zero, Now, imagine inscribed eylinders with increasing height, but decreasing ra- dius. Initially, the volumes would increase from zero, but then they would decrease to zero as the cylinders’ heights get close to the cone’s height. Intuitively, the vol- lume should pezk for some cylinder. Since the radius is squared in the volume for- ‘mula, it counts more than the height and we would expect r > hat the maximum. ‘The volume of the inscribed cylinder is V=arh From similar triangles, which gives hea-Gr ‘When we substitute this expression for fin the formula for V, we obtain V=ar(a-$r) = sa" ‘We wish to maximize V for rin the interval [0,). Now én ine afi anus = ra2 $1) Figure 5 Figure 6 — comem on Section 3.4 Practical Problems 169 this value for r in the equation connecting r and h, we find that h = a/3. In other words, the inscribed cylinder has greatest volume when its radius is two-thirds the radius of the cone’s base and its height is one-third the altitude of the cone, Suppose that a fish swims upstream with velocity» relative to ihe water and thatthe current of the river has velocity ~e (the negative sign ind cates that the current’ velocity sin the direction opposite that ofthe fish). The en- ergy expended in traveling a distance dup the river is directly proportional t0 the time required to travel the distance d and the cube ofthe velocity. What veloity 0 minimizes the energy expended n swimming this distance? SOLUTION Figure 5 illustrates the situation, Since the fish’s velocity up the stream (ie, relative to the banks of the stream) is » ~ v,, we have d = (v = v,)t, where ris the required time. Thus ¢ = d/(o — »,). Fora fixed value of », the energy required for the fish to travel the distance d is therefore diy e Ev ‘The domain for the function F is the open interval (v,, 2°). To find the value of v that minimizes the required energy we set £"(v) = O-and solve for v: (= w= VU) kd sarny ay wwe wap (0) = ka 0, ‘The only critical point in the interval (1), 00) is found by solving 2v ~ 30, 3 which leads to » = 521. The interval is open so there are no end points to check. ‘The sign of £"(v) depends entirely on the expression 20 ~ 3n,, since all the other expressions are positive. If v <> then 20 ~ 3; > 0 so Eis increasing to the right of ve then 20 ~ 3n, < 080 Eis decreasing to the 3 7% 1-yields a local minimum. Since this is the left of Jy. It > Thus. by the First Derivative Test,» = 5 only critical point on the interval (vp, °¢), this must give a global minimum. The velocity that minimizes the expended energy is therefore one and a half times the speed of the current. . EXAMPLE 5] A 6-foot-wide hallway makes a right-angle turn, What is the length of the longest thin rod that can be carried around the corner assuming you, ccannot tilt the rod? SOLUTION The rod that barely fits around the corner will touch the outside walls as well as the inside corner. As suggested in Figure 6, let a and b represent the lengths of the segments AB and BC, and let # denote the angles ZDBA and ZFCB. Consider the (wo similar right triangles, AADB and ABFC; these have hypotenuses a and 4, respectively. A little trigonometry applied to these angles ives 6 6 a= Fyn osee# and b= T= bese Note that the angle # determines the position of the rod. The total length of the rod. in Figure 6 is thus, 19) 6 sec 0 + besc8 ‘The domain for 0 is the open terval (0, 2/2). The derivative of L is 170. Chapter 3 Applications of the Derivative 1/0) = 6 see 0 tan 0 ~ 6 ese Ocot 0 _ me sino one) coed sin? @ sin’ @ — cos?@ sin? @ cos" @ 6 ‘Thus 1/(0) = 0 provided sin’ @ ~ cos’ = 0, This leads to sin @ = e084, The Scie only angle in (0, 7/2) for which sin@ = cos@ is the angle m/4 (see Figure 7). We again apply the First Derivative Test. 1F0 << /4, then sin 8 < cos @ (see Figure 7 again) so sin? # — cos! @ <0. Thus, (8) is decreasing on (0, m/4). It a/4 <8 < w/2then sin @ > cos so sin? # ~ cos’ > 0. Thus, L(A) is increasing oon (17/4, 2/2). By the First Derivative Test, @ = 2/4 yields a minimum. The prob- lem, however, asks for the longest rod that fits around the corner. As Figure 8 below indicates, we are actualy finding the smallest rod that satisfies the condi- : ” don in Figure 6 in other words, we are finding the smallest rod that does fit Rowe 7 around the corner. Therefore, the longest rod that does fit around the comer is Ln/d) = 6seen/8 + 6xe/4 = 12V2 = 1637 leet . _—— n cones nef Aenea vey en eet ‘wet Least Squares (Optional) There are a number of physical, economic, and social phenomena in which one variable is proportional to another. For example, Ltt Newton's Second Law says that the force F on an object of mass m is proportional to its acceleration a (F = ma). Hooke’s Law says that the force exerted by a spring is proportional to the distance it is stretched (FF = kx). (Hooke’s Law is often given as F = —kx, with the negative sign indicating that the force isin the direction opposite the stretch, For now, we will ignore the sign of the force.) Man= ufacturing costs are proportional to the number of units produced. The number of +,000000000000005-+ traffic accidents is proportional to the volume of traffic, These are modets, and in an experiment we will rarely find that the observed data fit the model exactly. Suppose that we observe the force exerted by a spring when itis stretched by x centimeters (Figure 9). For example, when we stretch the spring by 0.5 centi- Figure 9 meter (0.005 meter), we observe a force of 8 newtons, when we stretch the spring by 1.0 centimeter, we observe a force of 17 newtons,and so on, Figure 10 shows ad- ‘Spring steed by amount “sina inal ditional observations, and Figure 11 shows a plot of the ordered pairs (1 3. Sa where x, is the distance stretched and y, is the force exerted on the spring. A plot of nes ordered pairs like this is called a seatfer plot. so Let's generalize the problem to one in which we are given m points | (21, yi): (4 32),---» (4m Yo)- Our goal is to find a line through the origin that best thas {its these points. Before proceeding, we must introduce sigma (¥) notation Figure 10 ‘The symbol Sa, means the sum of the numbers ay, 4»,..-, dy. For example, Disance rete (meters) Section 3.4 Practical Problems 171 BeBe Fa and Says tame to + sae In the second ease, we multiply x, and yy; first and then sum. To find the line that best fits these m points, we must be specific about how we measure the fit, Our best-fit line through the origin is defined to be the one that izes the sum of the squared vertical deviations between (.x, 2) and the Tine y= bx. If (x, 1) iS a point in the data set, then (x, bx,) is the point on the line y= bx that is directly above or below (1x, :). The vertical deviation between (ss) and (x, bx,) is therefore y, — bx;, (See Figure 12.) The squared deviation is thus (yj — bx,)!, The problem is to find the value of b that minimizes the sum of these squared deviations If we define = So- on? then we must find the value of b that minimizes S.This is a minimization problem like the ones encountered before. Keep in mind, however. that the ordered pairs (45, yf = 1,2,.-., mare fixed; the variable in this problem is b, We proceed as before by finding dS/dé, setting the result equal 10 0, and solv ing for b. Since the derivative isa linear operator, we have aay : By = $6, — omg! od 2 a bx,) = S20 - oma( Lor - 6x9) ~23 aly, ~ bu) Setting this result equal to zero and solving yields 0= Duy 0 De which is always positive, There are no end points to check. Thus by the Second Derivative Test, we conclude that the line y = by, with b = S'x,y, / Sad is the best-fit line, in the sense of minimizing S, The line y = bx is called the least- squares line through the origin, EXAMPLE 6] Find the least-squares line through the origin for the spring data in Figure 10. 172. Chapter 3 Applications of the Derivative Distance stretched (mete) Figure 13, y Re * Works pe Figure 15, a: cu ra Figure 16 SOLUTION 0.005 +8 + 0.01017 + 0.015+22 + 0,020+32 + 0.025 +36 0.005" + 0.010? + 0.013" + 0.020" + 0.025 ~ 1512.7 ‘The least-squares line through the origin is therefore y = 1512.7x and is shown in Figure 13.The estimate of the spring constant is therefore k = 1512.7. . For most line-fitting problems, it is unreasonable to assume that the line pass- cs through the origin, A more reasonable assumption is that y is related to x by y = a + bx. In this case, however, the sum of squares is a function of both a and b ‘So we are faced with the problem of minimizing a function of two variables, a prob- lem we address in Chapter 12. Economic Applications (Optional) Consider a typical company, the ‘ABC Company. For simplicity, assume that ABC produces and markets a single product;it might be television sets,car batteries, or bars of soup. Iit sells x units of the product in a fixed period of time (¢.g., a year), it will be able to charge a price, p(x), for each unit, In other words, p(x) is the price required to attract a demand for.x units. The total revenue that ABC can expect is given by R(x) = xp(x), the ‘number of units times the price per unit To produce and market x units, ABC will have u total cost, C(x). Tis is nor- mally the sum of a fixed eost (office utilities, real estate taxes, and so on) plus @ variable cost, which depends dizectly on the number of units produced ‘The key concept fora company isthe total profit, P(x). tis just the difference between revenue and cost; that i, P(x) = R(x) ~ C(3) = xp(x) ~ (3) Generally. company seeks to maximize its total profit There isa feature that tends to distinguish problems in economics from those in the physial seiences. In most cases, ABC's produet will hein diserete units (you can’t make or sell §23 television sets of zr cat batteies)."Thus the functions R(x), (3), and P(x) are usually defined only for x = 0,1,2,... and, consequently, th graphs consist of dserete points (Figure 14). In order to make the tools of calculus available, we connect these points with a smooth curve (Figure 15), thereby pre- tending that R, Cand P ate nie differentiable functions. This illustrates an aspect of mathematical modeling that is almost always necessary. especially in economics. ‘To model a real-world problem, we must make some simplifying assumptions. This means that the answers we get are only approximations of the answers that we seek—one of the reasons economies isa less than perfect science. A well-known statistician once said: No model is accurate, but many models are useful. {A related problem for an economist is how to obtain formulas for the func- tions C(x) and p(x). In a simple case, C(x) might have the form C(x) = 10.000 + Sox: If so, $10,000 is the fixed est and $50x is the variable cost, based on a $50 direct, cost for each unit produced. Pethaps a more typical situation is C\(x) = 10,000 + 45x + 100VE Both cost functions are shown in Figure 16. ‘The cost function C(x) indicates that the cost of making an additional unit is the same regardless of how many units have been made. On the other hand, the cost function C(x) indicates that the cost of making additional units increases but at a decreasing rate. Thus, C\(x) allows for what economists call economies of seale. Selecting appropriate functions to model cost and price is a nontrivial task. ceasionally they can be inferred from basic assumptions. In other cases, a careful

You might also like