Tom K. answered 08/21/21
Knowledgeable and Friendly Math and Statistics Tutor
The gradient is
(4x, 6 - 2y), which has a 0 at (0, 3), which is in the interior. However, the second derivative is
4 0
0 -2
which tells us that this is a saddle point as eigenvalues are positive (4) and negative (-2)
We also can see that the minimum will be at (0, -4) and is 2(0)2 - (-4)2 +6(-4) = -40
However, we still will have to use the standard LaGrange procedure for the maximum. (This will also show us that the stated minimum is one of the candidates, s ptoog og yr
l(x,y) = 2x2 - y2 + 6y + l(x2+y2 - 16)
The gradient is (4x + 2xl, -2y + 6 + 2yl) = (2x(2 + l), -2y + 6 + 2yl)
We get the minimum mentioned above by setting 2x(2 + l) = 0 by letting x = 0, which leaves us free to vary l such that we get -2y + 6 + 2yl = 0 and are on the boundary. We can just plug in y = 4 and -4.
f(0, -4) is given above. f(0, 4) = 8. This will be a candidate for the minimum and maximum. Clearly, it is not the minimum. We could look at l if we chose, but that is not necessary.
2x(2 + l) = 0, also, when l = -2
Then, -2y + 6 + 2yl = 0
-2y + 6 + 2y(-2) = 0
-6y + 6 = 0
y = 1
As x2 + y2 = 16, x = ±√15.
f(-√15, 1) = f(√15, 1) = 35
This will be a maximum. (The other value at (0, 4) now makes sense, as within the boundary, this becomes a local minimum).
The absolute minimum is -40 at (0, -4)
The absolute maximum is 35 at (-√15, 1) and (√15, 1)