Critical points are points where the derivative of a function is equal to zero.
In multivariable calculus critical points are points where the partial derivatives of the function both equal zero.
To find the critical points of a function one finds the partial derivatives of the function, sets them equal to zero, and then solves the resulting system of equations.
Once the critical points have been obtained the second derivative test can be used to define each point as a minimum, maximum, or a saddle point.
See Multivariable Calculus Notes - Chapter 14 - Sections 7-8 for more information.
If your function is constrained within a region then your function must have an absolute minimum and maximum within that region.
The above theorem is called the extreme value theorem.
One can find these points via the following:
See Multivariable Calculus Notes - Chapter 14 - Sections 7-8 for more information and examples.