Beruflich Dokumente
Kultur Dokumente
(b) Use the second derivative test to classify the critical points as either min, max, or a saddle
point. If it is a max/min, is a global max/min? Briefly explain why/why not.
Answer. (a)
fx = 8x + 4y − 4 = 0
fy = 6y + 4x + 2 = 0
x = 1, y = −1.
2 = λyz
1 = λxz
1 = λxy
xyz = 1.
Eliminating λ we get y = 2x, z = y, 4x3 = 1 which yields (x, y, z) = 4−1/3 , 2 · 4−1/3 , 2 · 4−1/3 ,
1
• Boundary y = 4 − x, x ∈ [0, 4]:
2
f (x, 4 − x) = x2 + 2 (4 − x) − 2x − 4 (4 − x)
d
f (x, 4 − x) = 6x − 14 = 0 ⇐⇒ x = 7/3
dx
It has critical point at x = 7/3 (which corresponds to a min, with f = − 31 there) and a max
at (x, y) = (0, 4) (why?)
• In conclusion, the max occurs at the corner (0, 4) and the min occurs at the interior point
(1, 1) . The max value is 16, the min value is -3.
4. You want to make box with volume 16 m3 . The material for the sides costs half of what the material
for the top or bottom cost. Find the dimensions that will minimize the cost.
Answer. Solution 1 (the hard way): For the box of height z and whose other dimensions are
x, y, the cost is proportional to
f (x, y, z) = yz + xz + 2xy.
So we wish to minimize f (x, y, z) subject to the constraint xyz = 16. Setting this up using Lagrange
multiplier, we need to solve these four equations in four unknowns x, y, z, λ:
z + 2y = λyz
z + 2x = λxz
y + x = λxy
xyz = 16.
λ = 1, x, y = 2, z = 4.
Solution 2 (the easy way): By symmetry, the base of the box should be a square. Let x be one
of the dimensions of this square and let z be the height of the box. Then the cost is proportional
to f (x, z) = x2 + xz. We wish to minimize this f (x, z) subject to constraint x2 z = 16. Solve for
z = 16/x2 then f = x2 + 16/x, x ∈ [0, 4] . The min of this one-dimensional function occurs at
2x − 16/x2 = 0 or x3 = 8 or x = 2, z = 4.
5. The plane x + y + 2z = 2 intersects the surface z = x2 + y 2 . Find the point on the intersection of
these two surfaces which is (a) closest to z-axis and (b) furthest away from the z-axis.
Answer. We wish to extremize f (x, y) = x2 + y 2 (i.e. distance squared to the z-axis) subject to
two constraints g = x + y + 2z − 2 = 0, h = x2 + y 2 − z = 0. So we need to solve five equations:
∇f = λ∇g + µ∇h, g = h = 0 in five unknowns (x, y, z, λ, µ) :
2x = λ + µ2x
2y = λ + µ2y
0 = 2λ − µ
z = x2 + y 2
x + y + 2z = 2.
Eliminate λ, µ from the first three equations yields (1 + 4x) y = (1 + 4y) x. Together with the last
two, this yields two solutions
1 1 1
(x, y, z) = (−1, −1, 2) and (x, y, z) = , , .
2 2 2
Of course the former is the max of f and the latter is its min.