Solutions to Suggested Problems Section 7.1 33. Solving the equation x − y2 = z0 for x, we see that the level curves are parabolas (with horizontal axes of symmetry) of the form x = y2 + z0 . The x-section at x = 0 is the parabola z = −y2 ; the y-section at y = −2 is the line z = x − 4. (i) level curves (ii) graph of z=f(x,y) 4 2 2 0 y –2 1 –4 –6 –8 –2 –1 0 1 2 3 x 4 5 –10 –3 –2 –1 –1 y –2 0 1 2 3 1 5 4 3 2 1 x 0 –1 –2 35. Writing the equation 1+x2 +y2 = z0 in the form x2 +y2 = z0 −1, we see that the level curves are circles (centered at the origin), if z0 > 1. If z0 = 1, then the level curve is the point (0,0). If z0 < 0, then the level curve is the empty set. The x-section at x = 2 is the parabola z = y2 + 5; the y-section at y = −3 is the parabola z = x2 + 10. (i) level curves (ii) graph of z=f(x,y) 4 30 y 2 25 20 15 –4 –2 2 4 10 x 5 –2 4 2 0 x –2 –4 –4 2 2 –2 0 y –4 37. For z0 6= 0, the level curves are of the form xy = z0 . If z0 = 0 the level curve consists of the two coordinate axes. The x-section at x = 1 is the line z = y; the y-section at y = −1 is the line z = −x. (a) level curves (b) graph of z = f (x, y) 4 15 10 y 2 5 0 –5 –4 –2 0 2 4 –10 x –15 –4 –2 –2 x 0 2 4 –4 3 –4 –2 2 0 y 4 39. (a) When L = L0 , we have Q = Q(r, L0 ) = πr4 P/(8L0 v) = Kr4 , where K = πP/(8L0 v) (a positive constant). (b) When r = r0, we have Q = Q(r0 , L) = πr04 P/(8Lv) = K/L, where K = πr04 P/(8v) (a positive constant). (b) graph of Q = (a) graph of Q = Kr4 K L 40 2 35 1.5 30 25 1 20 15 0.5 10 5 0 0.2 0.4 0.6 0.8 1 0.2 r 0.4 0.6 L 4 0.8 1 41. (a) The revenue (in dollars) is given by the formula R(x, y) = 2, 000x + 3, 000y, where x is the number of units of product X sold and y is the number of units of product Y . (b) If we fix R at 60,000 we get the equation 2, 000x + 3, 000y = 60, 000 or y = − 23 x + 20. When we fix R at 120,000, we get the equation 2, 000x + 3, 000y = 120, 000 or y = − 23 x + 40. (top line is R=120,000, bottom line is R=60,000) 40 30 y 20 10 0 10 20 30 40 50 60 x 5 42. When the given production level is fixed at 15,000 units, we obtain the equation 15K 3/4 L1/4 = 15, 000, or K 3/4 L1/4 = 1, 000. Raising each side 12 to the 4th power gives K 3 L = 1012 , and solving for L gives L = 10K 3 . If Q = 15, 000 and K = 100, then L = 1012 1003 = 106 = 1, 000, 000 units. 5000 4000 3000 L 2000 1000 0 1000 2000 3000 4000 5000 K 59. In this case, (x0 , y0 , z0 ) = (4, −3, 3), a = 2, and b = 5. Therefore, the equation of the plane is z − 3 = 2(x − 4) + 5(y + 3), or z = 2x + 5y + 10. 63. Let (x0 , y0 , z0 ) = (1, 0, −1). Then z − (−1) = a(x − 1) + b(y − 0) or, z + 1 = a(x − 1) + by. To find a and b, we substitute (x, y, z) = (0, 1, 3) to get 3 + 1 = a(0 − 1) + b, or 4 = −a + b (x, y, z) = (1, −1, −2) to get −2 + 1 = a(1 − 1) + b(−1), or −1 = −b Thus, a = −3 and b = 1. The equation of the plane is z + 1 = −3(x − 1) + y, or z = −3x + y + 2. 66. z = z0 + ax + by (from equation 2), and a = 3, b = 5. So z = z0 + 3x + 5y, and ∆z = 3∆x + 5∆y. Setting ∆x = 0.1 and ∆y = −0.2, we find that the change in z is ∆z = 3(0.1) + 5(−0.2) = −0.7. 72. Let P(x, y) be the profit in selling x units of item X and y units of item Y . From the information given, we know that P(x, y) = 2x + 5y + c for some real number c, and P(100, 50) = −500. Therefore, −500 = P(100, 50) = 6 2 · 100 + 5 · 50 + c = 200 + 250 + c = 450 + c, so that c = −950. Thus, P(x, y) = 2x + 5y − 950, and P(100, 500) = 2(100) + 5(500) − 950 = 200 + 2, 500 − 950 = 1, 750 dollars. Section 7.2 3. ∂f ∂x = 3x2 y2 − 4xy, ∂∂yf = 2x3 y − 2x2 + 1. 5. ∂M ∂r rt = 1000tert , ∂M ∂t = 1000re . 7. ∂R ∂s = − 5ts2 , ∂R ∂t = 9. ∂l ∂x = 11. 2 ∂Q ∂K 10t s . ∂ 2 1 2x . (x − 5y + 3) = x2 −5y+3 x2 −5y+3 ∂x 0.4 −0.4 . = 4K −0.6 L0.6 , ∂Q ∂L = 6K L 13. ∂M r 360t r = 1000(1 + ) [ln(1 + )] · 360 ∂t 360 360 r 360t r = 360, 000(1 + ) ln(1 + ). 360 360 ∂M r 360t−1 1 = 1000 · 360t(1 + ) ( ) ∂r 360 360 r 360t−1 = 1000t(1 + ) . 360 15. ∂f ∂x = 2x + y, ∂∂xf (2, 1) = 2 · 2 + 1 = 5. 17. ∂f ∂x = 50x−1/2 y1/2 , ∂∂xf (25, 16) = 50 · 25−1/2 · 161/2 = 50 · 15 · 4 = 40. 7 21. (a) ∂T ∂x = 10x, ∂T ∂y = −10y. (b) If x = a, the x-section has the equation T = 5(a2 − y2 ) + 75 = −5y2 + (75 + 5a2 ). It is a parabola, opening downward, that achieves a maximum at y = 0. (c) If y = b, the y-section has the equation T = 5(x2 − b2 ) + 75 = 5x2 + (75 − 5b2 ). It is a parabola, opening upward, that achieves a minimum at x = 0. (d) If there were a local minimum (or maximum), both the x- and y-sections through the point would have a local minimum (or maximum), which cannot be the case. 23. The equation of the tangent plane has the form z = f (x0 , y0 ) + ∂f ∂f (x0 , y0 )(x − x0 ) + (x0 , y0 )(y − y0 ). ∂x ∂y Since ∂∂xf = 6x and ∂∂yf = 10y, we have ∂∂xf (1, 2) = 6, and ∂∂yf (1, 2) = 20. Also, f (1, 2) = 23. The equation of the tangent plane is given by z = 23 + 6(x − 1) + 20(y − 2) = −23 + 6x + 20y, and the linear approximation is f (x, y) ≈ −23 + 6x + 20y. 27. Since ∂f ∂y (4, 1) ∂f ∂x = 4x−1/2 y2 and ∂f ∂y = 16x1/2 y, we have ∂f ∂x (4, 1) = 2 and = 32. Also, f (4, 1) = 16. The equation of the tangent plane is given by z = 16 + 2(x − 4) + 32(y − 1) = −24 + 2x + 32y, and the linear approximation is f (x, y) ≈ −24 + 2x + 32y. 31. z = −2 + 2(x + 3) − 78 (y − 5) = 2x − 87 y + 68 7 . The linear approximation 8 68 is f (x, y) ≈ 2x − 7 y + 7 . 33. The volume of a rectangular box whose height is x inches and whose base is a square with each side y inches long equals V (x, y) = xy2 . (a) V (6, 4) = 6 · 42 = 96 cubic inches. 8 ∂V 2 (b) Since ∂V ∂x = y and ∂y = 2xy, we obtain The linear approximation formula gives ∂V ∂V ∂x (6, 4) = 16 and ∂y (6, 4) = 48. ∂V ∂V (6, 4) · (6.02 − 6) + (6, 4) · (4.01 − 4) ∂x ∂y = 16 · (0.02) + 48 · (0.01) = 0.8. V (6.02, 4.01) −V (6, 4) ≈ Section 7.3 1. We must solve the system of equations ∂f = 2x − 2 = 0 ∂x ∂f = 2y + 4 = 0 ∂y From the first equation we get x = 1, and from the second equation we get y = −2. Thus, (1, −2) is the only critical point. 3. Setting equations ∂f ∂x = −5y − 4x = 0 and ∂f ∂y = 2y − 5x = 0 we obtain the system of 4x + 5y = 0 5x − 2y = 0, whose unique solution is x = 0, y = 0. Thus, (0, 0) is the only critical point. 5. Setting ∂∂xf = 3y − 3x2 = 0 and of equations ∂f ∂y = 3x − 3y2 = 0 we obtain the system y = x2 x = y2 , which leads to y = y4 , whose solutions are y = 0 and ,y = 1. If y = 0, then x = 0, and if y = 1 then x = 1. Thus, (0, 0) and (1, 1) are the critical points. 9 9. Set ∂∂xf = 2x = 0 and ∂∂yf = 2yey = 0. The only solution is x = 0 and y = 0. Thus, (0, 0) is the only critical point. 2 11. Setting ∂g ∂x = 2x − y = 0 and of equations ∂g ∂y = −x + 2y + 3 = 0, we obtain the system 2x − y = 0 x − 2y = 3 This system has the unique solution x = −1 and y = −2. Thus, (−1, −2) 2 is the only critical point. The second order partial derivatives are ∂∂xg2 = 2, ∂2 g ∂x∂y ∂ g 2 = −1 and ∂∂yg2 = 2, and D(x, y) = ∂∂xg2 · ∂∂yg2 −( ∂x∂y ) = 2·2−(−1)2 = 3 > 2 2 0. Since D(x, y) > 0 and ∂2 g ∂x2 2 2 = 2 > 0, g has a local minimum at (−1, −2). 15. Setting ∂∂xf = 3x2 − 3y = 0 and of equations ∂f ∂y = 3y2 − 3x = 0, we obtain the system y = x2 x = y2 whose solutions are x = 0, y = 0 and x = 1, y = 1. Thus, (0, 0) and (1, 1) 2 are the critical points. The second order partial derivatives are ∂∂x2f = 6x, ∂2 f ∂x∂y = −3 and ∂2 f ∂2 f ∂y2 = 6y, and so D(x, y) = 36xy − 9. Thus D(1, 1) = 27 > 0 and ∂x2 (1, 1) = 6 > 0, so f has a local minimum at (1, 1); and D(0, 0) = −9 < 0, so that f has a saddle point at (0, 0). 29. Set ∂f ∂x = 4x − 12 = 0 and ∂f ∂y = 6y − 6 = 0. From the first equation, we ∂2 f ∂2 f = 4, 2 ∂x∂y = 0 ∂x ∂2 f > 0. Therefore, ∂x2 get x = 3, and from the second we get y = 1. In addition, and ∂∂y2f = 6, and so D(x, y) = 24. Thus D(x, y) > 0 and f has a local minimum at (3, 1), and since f is a quadratic function the local minimum is also a global one. 2 10 31. Set ∂f ∂x = 6 − 6x = 0 and ∂f ∂y = 4 − 8y = 0. The only solution is x = 1 ∂2 f = −6, ∂x2 2 and ∂∂x2f < 0. and y = 1/2. Thus, (1, 12 ) is the only critical point. In addition, ∂2 f ∂x∂y = 0 and ∂2 f ∂y2 = −8, and so D(x, y) = 48. Thus D(x, y) > 0 Therefore, f has a local maximum at (1, 12 ), and since f is a quadratic function the local maximum is also a global one. 35. Setting ∂P ∂x = 8 − 0.02x − 0.01y = 0 and we obtain the system of equations ∂P ∂y = 11 − 0.01x − 0.04y = 0, 2x + y = 800 x + 4y = 1100 The system has the solution x = 300 and y = 200. In addition, ∂2 P ∂2 P ∂y2 ∂2 P ∂x2 = −0.02 < 0, ∂x∂y = −0.01 and = −0.04, and so D(x, y) = 0.0007 > 0. Thus, the profit is maximized if x = 300 and y = 200, with the maximum profit being P(300, 200) = 2400+2200−900−0.01(3002 +300·200+2·2002 ) = 1400. 39. (a) The weekly revenue of the toy store is 1 R(x, y) = x(50 − 3x + 2y) + y(6 + x − y) 2 5 = 50x + 6y + xy − 3x2 − y2 . 2 5 (b) Setting ∂R ∂x = 50 + 2 y − 6x = 0 and system of equations ∂R ∂y = 6 + 52 x − 2y = 0, we obtain the 12x − 5y = 100 5x − 4y = −12 whose solution is x = 20 and y = 28. In addition, ∂2 R ∂y2 = −2. Since D(x, y) = −6(−2) − ( 52 )2 = 11 23 4 ∂2 R ∂x2 = −6, > 0 and ∂2 R ∂x2 ∂2 R ∂x∂y = 5 2 and = −6 < 0, the revenue function has a local maximum at x = 20 and y = 28. Because the function is quadratic, it also has a global maximum at that point. (c) The weekly profit of the toy store is P(x, y) = R(x, y) − 9(50 − 3x + 2y) − 14(6 + 0.5x − y) 5 = 70x + 2y + xy − 3x2 − y2 − 534. 2 5 Setting ∂P ∂x = 70 + 2 y − 6x = 0 and system of equations ∂P ∂y = 2 + 52 x − 2y = 0, we obtain the 12x − 5y = 140 5x − 4y = −4 748 whose solution is x = 580 23 ≈ 25.22 and y = 23 ≈ 32.52. Applying again the second derivative test, we see that those are the prices the store should charge for the standard and deluxe games, respectively, to maximize its profit. 41. stove is placed at the point (x, y), the total distance is r(x, y) = p (a) If thep ∂r ∂r 2 2 x + y + (x − 3)2 + (y − 5)2 . Setting ∂x = 0 and ∂y = 0 leads to the system of equations x p x2 + y2 y p x2 + y2 = p = p 3−x (x − 3)2 + (y − 5)2 5−y (x − 3)2 + (y − 5)2 Assuming y 6= 0 and y 6= 5 (which must be true for the second equation to be valid), we can divide the first equation by the second to get x 3−x = y 5−y which reduces to x(5 − y) = y(3 − x) and further simplifies to 5x = 3y, or y = 35 x. Conversely, if y = 53 x and 0 < x < 3, then ∂r x x−3 = p +p ∂x (x − 3)2 + (y − 5)2 x2 + y2 3x 3(x − 3) 3 3 = √ +p = √ −√ =0 2 2 34 34 34(x − 3) 34x 12 and similarly ∂r ∂y = 0. (b) We have seen that any pair (x, y) satisfying y = 35 x and 0 < x < 3 is a critical point. These are the points of the line segment between (0, 0) and (3, 5). A geometric argument shows that choosing any such point minimizes the total distance. In fact, this exercise is an analytic verification of the wellknown geometric principle that the shortest distance between two points is along the straight line between them. 47. Setting ∂E ∂a ∂E ∂b = 2(a + b − 2) + 2(2a + b − 2) · 2 + 2(3a + b − 4) · 3 = 28a + 12b − 36 = 0 = 2(a + b − 2) + 2(2a + b − 2) + 2(3a + b − 4) = 12a + 6b − 16 = 0 we obtain the system of equations 7a + 3b = 9 6a + 3b = 8. This system has the solution a = 1 and b = 23 . In addition, ∂2 E = 28, ∂a2 ∂2 E = 12, ∂a∂b ∂2 E = 6. ∂b2 Therefore, D(a, b) = 28 · 6 − 122 = 24 > 0 and ∂∂aE2 = 28 > 0. Since E is a quadratic function, it has a global minimum at (1, 23 ). 2 Section 7.5 1. Given f (x, y) = x2 + y2 with constraint g(x, y) = 2x + 6y − 2000 = 0, we ∂f ∂g set ∂∂xf = λ ∂g ∂x and ∂y = λ ∂y , leading to the system of equations 2x = 2λ 2y = 6λ 2x + 6y = 2000 From the first two equations we obtain x = λ and y = 3λ, so that y = 3x. Substituting that into the third equation gives 2x + 18x = 2000 or x = 100, 13 and it follows that y = 300. Therefore, if there is a local minimum or maximum, subject to the constraint, it must occur at (100, 300). From geometric considerations, we can conclude that f (x, y) does, in fact, have a minimum at (100, 300), subject to the constraint. For the level sets of f are circles centered at the origin, and the one tangent to the constraint line g(x, y) = 0 at (100, 300) has the smalles radius of all those that intersect the line. 5. Given f (x, y) = (x − 1)2 + (y − 2)2 − 4 and g(x, y) = 3x + 5y − 47, we ∂F ∂F define F(x, y, λ) = f (x, y) − λg(x, y) and set ∂F ∂x = 0, ∂y = 0, and ∂λ = 0. These are equivalent to system of equations ∂f ∂x ∂f ∂g = λ ∂g ∂x , ∂y = λ ∂y , and g(x, y) = 0, and lead to the 2(x − 1) = 3λ 2(y − 2) = 5λ 3x + 5y = 47. From the first two equations, we obtain 10(x − 1) = 6(y − 2), which reduces to 5x−3y = −1. Combining that with the third equation leads to the solution x = 4, y = 7. We next apply the 2nd derivative test. The partial derivatives are: ∂F ∂2 F ∂2 F ∂2 F ∂F = 2(x−1)−3λ, = 2(y−2)−5λ, = 2, = 0, = 2. ∂x ∂y ∂x2 ∂x∂y ∂y2 Since D(x, y) = 2 · 2 − 02 = 4 > 0 and ∂∂xF2 = 2 > 0, f (x, y) has a local minimum at (4, 7), subject to the constraint g(x, y) = 0. 2 We can also reach the same conclusion from geometric considerations similar to that of exercise 1 involving the level sets, which are circles centered at (1, 2). In fact, we reach the stronger conclusion that there is a global minimum at (4, 7). 7. Given f (x, y) = xy with the constraint g(x, y) = x2 + y2 − 1 = 0, we set 14 ∂f ∂x = λ ∂g ∂x and ∂f ∂y = λ ∂g ∂y , leading to the system of equations y = 2xλ x = 2yλ 2 x + y2 = 1. By multiplying the first equation by y and the second by x, we√obtain y2 = 2xyλ = x2 . That reduces the third equation√to 2x2√= 1, or x =√± 2/2.√Thus, there √ are four √ possible√extreme√points: ( 2/2, 2/2), (− 2/2, − 2/2), (− 2/2, 2/2) and ( 2/2, − 2/2). The constraint curve is the circle of radius 1 about the origin, and the level curves of f are of the form xy = c. If c > 0, each curve has two branches, one in the first quadrant and the other in the third. Among those curves that intersect the constraint√circle, is 1/2, occuring at the √the largest value √ of c √ two tangency points ( 2/2, 2/2), and (− 2/2, − 2/2). Thus, f has a global maximum, subject to the constraint, at each of those points. Similarly,√ f has√a global minimum, subject to the constraint, at each √ √ of the points (− 2/2, 2/2) and ( 2/2, − 2/2), where c = −1/2. 9. Given f (x, y) = x2 + y2 with the constraint g(x, y) = xy − 1 = 0, we set ∂f ∂f ∂g ∂g ∂x = λ ∂x and ∂y = λ ∂y , leading to the system of equations 2x = λy 2y = λx xy = 1. Multiplying the first equation by x and the second by y, we obtain 2x2 = λxy = 2y2 , so that y = ±x. However, y 6= −x, otherwise the third equation would become −x2 = 1. Therefore, y = x, and by substituting that into the third equation we obtain x2 = 1, or x = ±1. Thus, the possible extreme points are : (1, 1), and (−1, −1). By considering the relation of the level curves of f to the constraint curve as we have done in the previous exercises, we conclude that, subject to the constraint, f has a global minimum at each of the points (1, 1) and (−1, −1), with value f (±1, ±1) = 2. 15 13. With f (x, y) = x2 + y2 and g(x, y) = 3x + y − 1, we set ∂f ∂g ∂y = λ ∂y , leading to the system of equations ∂f ∂x = λ ∂g ∂x and 2x = 3λ 2y = λ 3x + y = 1. From the first two equations we get y = 31 x, and substituting that into the 3 1 3 . Thus, the only extreme point is ( 10 , 10 ). third equation gives x = 10 An examination of the level curves, similar to that in exercise 1, shows that f has a global minimum, subject to the constraint g(x, y) = 0, at the point 3 1 ( 10 , 10 ), which is therefore the point on the line closest to the origin. As an alternative, we can apply the second derivative test to F(x, y, λ) = 3 1 1 , 10 , 5 ). (x2 + y2 ) − λ(3x + y − 1) at the point (x, y, λ) = ( 10 23. We need to maximize P(x, y) subject to the constraint Q(x, y) = x + y − ∂Q ∂Q ∂P 100 = 0. Setting ∂P ∂x = λ ∂x and ∂y = λ ∂y , we obtain the system of equations 10 − 0.2x = λ 20 − 0.2y = λ x + y = 100, whose unique solution is x = 25, y = 75, and λ = 5. Applying the second derivative test with F(x, y, λ) = 10x + 20y − 0.1(x2 + y2 ) − λ(x + y − 100), we find that ∂F ∂F ∂2 F ∂2 F ∂2 F = 10−0.2x−λ, = 20−0.2y−λ, = −0.2, = 0, = −0.2. ∂x ∂y ∂x2 ∂x∂y ∂y2 Then D(x, y) = (−0.2) · (−0.2) − 02 = 0.04 > 0 and ∂∂xF2 = −0.2 < 0, from which we conclude that P(x, y) has a maximum at (25, 75). When the company manufactures 25 units of X and 75 units of Y , it achieves its maximum profit P(25, 75) = 1, 125. 2 25. We need to maximize Q(K, L) under the constraint 70K +65L = 4, 000, 000. ∂g ∂Q Letting g(K, L) = 70K + 65L − 4, 000, 000 and setting ∂Q ∂K = λ ∂K and ∂L = 16 ∂g , we obtain the system of equations λ ∂L 14K −0.65 L0.65 = 70λ 26K 0.35 L−0.35 = 65λ 70K + 65L = 4, 000, 000. whose solution is K = 20, 000, and L = 40, 000. Thus, the combination of 20,000 units of capital and 40,000 units of labor gives the maximum output. (Note: You must still verify that this is indeed a maximum). 30. (a) We want to optimize S(x, y) = x3/4 y1/4 under the constraint x+y = 4. ∂T ∂S ∂T Write T (x, y) = x + y − 4 and set ∂S ∂x = λ ∂x and ∂y = λ ∂y . We obtain the system of equations 3 −1/4 1/4 x y = λ 4 1 3/4 −3/4 x y = λ 4 x + y = 4. whose solution is x = 3, y = 1. (b) There is a maximum at (x, y) = (3, 1), subject to the assumptions that x ≥ 0, y ≥ 0 and x + y = 4. To see why, observe that S(x, y) ≥ 0 all along the segment of the line x + y = 4 connecting the points (0, 4) and (4, 0), and S(0, 4) = S(4, 0) = 0. Therefore, there must be a maximum at some point of the segment, and the only candidate is (3, 1). (c) This crepuscular mammal should spend 3 of the dawn hours in foraging and 1 in defending. In other words, 3/4 of the dawn time should be spent foraging and 1/4 in defending. 17