18.5. LAGRANGE MULTIPLIERS 339

Hence y2−2λxy = 0 and x(x+1)−2λxy = 0 so y2 = x(x+1). Therefore from the con-straint, x2 + x(x+1) = 1 and the solution is x = −1,x = 1

2 . Then the candidates for a

solution are (−1,0) ,(

12 ,√

32

),(

12 ,−√

32

). Then

f (−1,0) = 0, f

(12,

√3

2

)=

3√

34

, f

(12,−√

32

)=−3

√3

4.

It follows the maximum value of this function is 3√

34 and it occurs at

(12 ,√

32

). The mini-

mum value is − 3√

34 and it occurs at

(12 ,−

√3

2

).

Example 18.5.3 Find candidates for the maximum and minimum values of the functionf (x,y) = xy− x2 on the set

{(x,y) : x2 +2xy+ y2 ≤ 4

}.

First, the only point where ∇ f equals zero is (x,y) = (0,0) and this is in the desired set.In fact it is an interior point of this set. This takes care of the interior points. What aboutthose on the boundary x2+2xy+y2 = 4? The problem is to maximize xy−x2 subject to theconstraint, x2 + 2xy+ y2 = 4. The Lagrangian is xy− x2−λ

(x2 +2xy+ y2−4

)and this

yields the following system.

y−2x−λ (2x+2y) = 0x−2λ (x+ y) = 0x2 +2xy+ y2 = 4

From the first two equations,

(2+2λ )x− (1−2λ )y = 0, (1−2λ )x−2λy = 0

Since not both x and y equal zero, it follows

det

(2+2λ 2λ −11−2λ −2λ

)= 0

which yields λ = 1/8. Therefore, y = 3x. From the constraint equation x2 + 2x(3x) +(3x)2 = 4 and so x = 1

2 or − 12 . Now since y = 3x, the points of interest on the boundary of

this set are (12,

32

), and

(−1

2,−3

2

). (18.1)

f(

12,

32

)=

(12

)(32

)−(

12

)2

=12

f(−1

2,−3

2

)=

(−1

2

)(−3

2

)−(−1

2

)2

=12

Thus the candidates for maximum and minimum are( 1

2 ,32

),(0,0), and

(− 1

2 ,−32

). There-

fore it appears that (0,0) yields a minimum and either( 1

2 ,32

)or(− 1

2 ,−32

)yields a max-

imum. However, this is a little misleading. How do you even know a maximum or aminimum exists? The set x2 + 2xy+ y2 ≤ 4 is an unbounded set which lies between the

18.5. LAGRANGE MULTIPLIERS 339Hence y* — 2Axy = 0 and x(x+1)—2Axy =0 so y? =x(x+1). Therefore from the con-straint, x* +.x(x+1) = 1 and the solution is x = —1,x = 5. Then the candidates for asolution are (—1,0), (4. 2) . (4. =). Then2? 2rn =0./(5.%8) 8 ( *) 3V33v3 v3It follows the maximum value of this function is “}~ and it occurs at (3. #8). The mini-mum value is 33 and it occurs at (3. -¥).Example 18.5.3 Find candidates for the maximum and minimum values of the functionf (x,y) = xy —x* on the set { (x,y) x2 4+ 2xy+y? < 4}.First, the only point where V f equals zero is (x,y) = (0,0) and this is in the desired set.In fact it is an interior point of this set. This takes care of the interior points. What aboutthose on the boundary x? +2xy+y? = 4? The problem is to maximize xy — x” subject to theconstraint, x* + 2xy +y? = 4. The Lagrangian is xy — x — A (x + 2xy+y? —4) and thisyields the following system.y—2x—A (2x+2y) =0x—2A (x+y) =0x +2xy+y?=4From the first two equations,(2+2A)x—(1—24)y =0, (1-2A)x-—2Ay =0Since not both x and y equal zero, it follows242A 2-1det ~0© ( 1-22 2awhich yields A = 1/8. Therefore, y = 3x. From the constraint equation x? + 2x (3x) +(3x) =4andsox= 5 or —5. Now since y = 3x, the points of interest on the boundary of_ (23)-m ($3) vs(3)-()G)-@)-(99)-) GG)Thus the candidates for maximum and minimum are (5,3) ,(0,0), and (—5,—3). There-fore it appears that (0,0) yields a minimum and either (5,3) or (—3,—3) yields a max-imum. However, this is a little misleading. How do you even know a maximum or aminimum exists? The set x* + 2xy + y* <4 is an unbounded set which lies between thewn NI