Previous | Next --- Slide 21 of 32
Back to Lecture Thumbnails

In a situation where we might have infinitely many optimal values for x, could we potentially add another constraint(s) that would guide us towards a unique x?


For the picture of the positive semidefinite case, is the value of the objective at the bottom of the cylinder all the same value? I kinda see it as slanting downwards on the right so it seems like we can continue going towards the right to get a better optimum.


What will be an example that satisfies this convex quadratic objective in real life?