I have a question asking to find the local max/min/saddle point of $f(x,y)=x^2+xy+y^2+y$

I have:







I know if D is more than zero then there should be extrema, and I had found (1/5,-2/5), (1,-2) as critical points, but with no way to test critical points by plugging them into $f_{xx}$ since it's a constant.... I'm not really sure where to go from here. They''re simple derivatives and I don't see where I messed up on them anywhere so I'm not sure what else to try.

  • $\begingroup$ Here I see $f_y = x + 2y + 1$. $\endgroup$ – Rolf Hoyer Oct 15 '16 at 4:06
  • $\begingroup$ Yea, typo, sorry. That's what I have on my paper. $\endgroup$ – windy401 Oct 15 '16 at 4:08
  • 1
    $\begingroup$ Just because $f_{xx}(x,y) =2$ doesn't mean we can't evaluate it at points! $f_{xx}(1/5, -2/5) = 2$ $\endgroup$ – erfink Oct 15 '16 at 4:25
  • $\begingroup$ Yea but every point would then give the same result, fxx > 0, so local min. So every critical point is a local min? The book is just saying the answer is that there is a local min at $x=1/3$ and $y=-2/3$. Still trying to figure out how they solved for that critical point.... Unfortunately I'm not good at solving for the zeros needed....Sigh. $\endgroup$ – windy401 Oct 15 '16 at 4:48

To get the critical points we need to solve $(f_x,f_y)=(0,0)$ for $(x,y)$. This gives two equations: (1) $2x+y=0$ and (2) $x+2y+1=0$. (1) is equivalent to $y=-2x$, and plugging this into (2) gives $x+2(-2x)+1=0$, so $3x=1$ and hence $x=1/3$. But then $y=-2x=-2/3$, showing that $(1/3,-2/3)$ is the only critical point.

To see if this is a local max/min, you need to plug this into the Hessian, which is the (constant) matrix $$\begin{pmatrix}2&1\\1&2\end{pmatrix}.$$ Since this matrix is positive definite we have a local min. at $(1/3,-2/3)$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.