How to check if three dimensional surface has endpoints

141 Views Asked by At

A big part of my calculus class is using Lagrange multipliers to find max/min values of a given function subject to some constraint. One thing I'm struggling with however, is that the endpoints of your given constraint curve are supposed to be considered as potential maxima/minima. This is trivial for some curves, and quite easy to see doesn't exist for constraints like spheres as they evidently have no endpoints, but there are some curves that I cannot visualize endpoints for.

For example, the constraint $z^{2} = xy +1$. I have a hard time visualizing it, and drawing it out is time consuming. My question is, how can I check for whether this kind of curve indeed has endpoints or not as simply as possible?

1

There are 1 best solutions below

9
On

The function can be written as

$$z=\pm\sqrt{xy+1}.$$

Obviously, if $xy<-1$ then there is no real $z$. Note that $xy<0$ if the signs are different. This means that our function has limitations only on the second and the fourth quadrant.

For a positive $x$ we have to find those negative $y$'s for which $xy<-1$. So, we have no solution if $$y<-\frac1x.$$

If $x<0$ then $y$ has to be positive and we don't have a solutions if $$y>-\frac1x.$$ (The inequality turned around because we divided by a negative number.)

The following figure depicts in red the domain where there is no solution, and in green where there is solution:

enter image description here

Alpha plots well depicts the situation. For $+\sqrt{xy+1}$ we have

enter image description here

And for $-\sqrt{xy+1}$ my Alpha PRO gives

enter image description here

The moral of my answer was supposed to be that we did not have to use Alpha, since elementary calculations showed the border lines.