i I i CHAPTER 20. OPTIMIZATION: FUNCTIC 'S OF SEVERAL VARIABLES ‘We will not dwell on the interpretation of these second-order derivatives. However, a few points should be made. The pure partial derivatives f,, and fy convey information about the concavity of a function (just as the second derivati does for single-variable functions). Specifically, fz» offers information about the concavity of traces which are parallel to the xz plane. Similarly, fyy provides information about the concavity of traces which are parallel to the yz plane. ‘The interpretation of the mixed partial derivatives fuy and fy is less intuitive than with the pure partial derivatives. However, they will be significant in the next section, Section 20.2 Follow-up Exercises In Exercises 1-40, determine f, and f,. 1 fz) = 32" 10)? 2 flay) =—2x4 By 3 flz,y) = 10x? + 2ay~ 6)? 4 fy) =x2— 29 5 fay) say 6 f(x,y) = 2539" 7 flx,3) = 6x2 — xy + 309? 8 fle. y) = 203? + Taty? + 5y? 9 flay) =~ 4/9" 10 f(s, 9) = ~2s/39? 11/69) = (2 — By) r+ ay") 12 fle, = (1/aN(9? + 39°) 13 fl) = (ey) 14 flz,9) = (a? +392)" 15 fix) =H 16 f(z,9)=—20Ale+ 17 f(y) =Inx ty) 18 f(s,9) =e 19 f(x,y) =e 20 f(z, 9) =e Iny 21 f(x,y) = 324 Bry? 22 f(x,y) = 10x? ~ 25xy + 30? 23 f(x, y)=—32" + Biy* + y* 24 f(x,y) = 15y*— Sy? +25 25 fla, 9) = 224/89? 26 fls,y) =~a/y? 28 f(x,y) = Bx—y*)/ax* +1) 30 f(x, 9) = Sy?/(2? ~ 5y) 82 f(s, y) = (Bx — 3y)* 83 fis y) VFB? 84 flx,) = VEE Oe 5 lx, 9) = Ina" ~y*) 36 f(s, 3) = Ing ~ 4 +) 87 fix, s)mer 38 flay) 40 f(s,9) =e In In Exercises 4160, find all second-order partial derivatives. 41 flx, 9) = 822+ 5y9+ 10 42 f(x,y) Bet dy? +2y 43 fz, ) = 5x*— Bay + By" 44 flz,9) = 2028 + 10sy + 6y? 45 fix.) =2'y* 46 f(z, 9) =—102y? AT fly) =e 48 f(z,y) =In(z +9) 49 fle) =07 50 fls,) 51 fly) = (9 82 f(z, 9) 58 f(x,y) = x49? 54 f(3,9) 85 f(y) 86 flz,9) 87 flx,9) 88 flx,y) = Very 59 f(x,y) = In xy" 60 f(x,y) =In Say? 61 Given f(x, ») = 1002* + 200)? — 10xy: (@) Determine /(10, 20).20.3. OPTIMIZATION OF BIVARIATE FUNCTIONS 947 (®) Using partial derivatives, estimate the change expected in f(x, »)ifx increases by 1 cont. (©) Compare the actual change with the estimated change. (@) Repeat parts b and c, assuming a possible increase in y of 1 unit. 62 Given f(x, y) = 20x* ~ 309? + 102%y: (@) Determine f(20, 10). (®) Using partial derivatives, estimate the change expected in f(x, )ifx increases by 1 nit, (©) Compare the actual change with the estimated change. (@) Repeat parts b and c, assuming a possible increase in y of 1 unit. 63 A firm estimates that the number of unitsit sells each year's a function ofthe advertis- ing expenditures for TV and radio. The function expressing this relationship is 2,000: + 6,000y —20x* — 10)" — 50ay where z equals the number of units sold, x equals the amount spent on TV advertising, {andy equals the amount spent on radio advertising (the latter twovarables expressed in $1,000s). The firm is presently allocating $50,000 to TV and $30,000 to radio. {o} What are annual sales expected to equal? (b) Using partial derivatives, estimate the effect on annual sales if an additional $1,000 is allocated to TV. (©) Using partial derivatives estimate the effect on annual sales ian addtional $1,000 in allocated to radio. (a) Where dos it seem thatthe $1,000 is better spent? 64 Given the demand function 1 MP4, Pe) = 25,000 ~ 0.1p} - 0.5pf (a) Determine the partial derivatives f, and f,,. (8) fp, = 20 and p, = 10, evaluate f,, and f,, and interpret their meaning. (©) How are the two products interrelated? 65 Given the demand function = (Pas Pay Ba) = 2507000 — O.6pf + p3 + Ops () Determine the partial derivatives ff. af. (8) py = 20, p= 10, and py = 20, evelae the pa rennin (o) How are the thee products interrelated? derivatives and interpret their 20.3 OPTIMIZATION OF BIVARIATE FUNCTIONS ‘The process of finding optimum values of bivariate functions is very similar to that, used for single-variable functions. This section discusses the process. Critical Points ‘As with single-variable functions, we will have a particular interest in identifying relative maximum and minimum points on the surface representing a function203 OPTIMIZATION OF BIVARIATE FUNCTIONS 959 fn? fy=0 | fom-8+U f=0 Evaluation of (0, 0,0): | (0,0) = (~2)(~6(0) + 24] — 08 =(-2)(24) -0 = <0 ‘Therefore, a saddle point occurs at (0, 0,0). Evaluation of (0, 8, 256): D(O,8) = (—2)[-6(8) + 24] - 0" 2-2 =48>0 With D> 0, the critical point is either relative maximum or a relative minimum. The values of the two pure partial derivatives at the critical point are (0, 8)=—2 and fol0,8) = ~6(8) + 24=—24. Since both are negative, we conclude thet a relative maxi- ‘mum occurs on the graph off at (0,8, 256). Figure 20.15 s sketch of f which provides two different views of the surface. Qa Section 20.3 Follow-up Exercises For the following functions, determine the location and nature of all critical points. 1 f(x,y) = 408 y+ 80x + 20y~ 10 2 fiz. y) = 22 +3y— Sx By? + By 3 f(x,y) = 35/3 — 5x3/2-+ 3y*— 129 4 f(x)9) = 824+ aay + 44x ~ 10? + 12y 5 f(z.y) = 3x*— day + By? + 8x — 1Ty +5 6 f(x,9) = 28+ 6x— Ly +99 +5 7 f(x, =28+ y*~ e+ 6y +10 B f(z, y) = 2? — Bay t By +z — 16y +22 9 f(x,y) = 2! +9! — Say 10 f(z, 9) =x +99 + Bay 11 f(y) =x + In x +9210, 2>0 12 f(z, 9) = 25x — 2bxe™? ~ 50y~ x? 18 flz,y) = x3/2-+ 293 — 202 — 40y + 100 14 f(x, 9) = ~33"~ 2y* + 45x — 809 — 50 15 f(x,y) = 322+ 9? + Bay ~ 60x ~ 329 +200 16 f(z 9) = x2 $y ty? 8 17 f(x,y) =3y- =" 18 f(x,y) = Be" day + 9y2-4-8x—1Ty+5 19 f(x) =—28" + Bay — 8 + 4x ~ 6y +10 20 f(x,y) = x2 + Dey + By? + Be + 10y— 20 21 f(x,y) = 2 y?— Bet hy +25960 CHAPTER 20. OPTIMIZATION: FUNCTIONS OF SEVERAL VARIABLES. fu) “ Figure 20.15 Relative maximum and saddle point on fx, )=— x" — y° + 12yt. 22 fls,y)=2x* +$y*—x— Ly +15 23 flx,)) 28+ y+ Gr +6 24 fx,9) = Ary 9-9" 2B flz,9) = 38+ y*— 32+ 6y +10