Tom K. answered 11/03/20
Knowledgeable and Friendly Math and Statistics Tutor
f(x,y) = 5y^2 + xy + y + 8x +6
The gradient is
(y + 8 10y + x + 1)
The Hessian is
0 1
1 10
There is only one critical point
From the gradient, from the first term, we see that y = -8. From the second term, 10y + x + 1 = 0
10(-8) + x + 1 = 0
x = 79
(79, -8) is our critical point
Note that no x or y appears in the Hessian.
The eigenvalues are solved from λ2 - 10λ - 1 = 0 At this point, we know that we have a positive and negative eigenvalue, so this is a saddle point. (We can solve the quadratic if you like; λ = 5 ± √26; the 0 as the first element of the Hessian makes it trivial to the see that the unnormalized eigenvectors are
(1 5 + √26) and (1 5 - √26)
If you move in the direction of the eigenvectors from (79, -8), you can verify that this is a saddlepoint. (moving in the direction of the first eigenvector, you will see an increase as one moves away from the critical point; moving in the direction of the second eigenvector, you will see a decrease).