1: \begin{abstract}

2: This paper studies generalized semi-infinite programs (GSIPs)

3: given by polynomials. We propose a hierarchy of

4: polynomial optimization relaxations to solve them.

5: They are based on Lagrange multiplier expressions and polynomial extensions.

6: Moment-SOS relaxations are applied to solve the polynomial optimization.

7: The convergence of this hierarchy is shown under certain conditions.

8: In particular, the classical semi-infinite programs (SIPs)

9: can be solved as a special case of GSIPs.

10: We also study GSIPs that have convex infinity constraints

11: and show that they can be solved exactly by

12: a single polynomial optimization relaxation.

13: The computational efficiency is demonstrated

14: by extensive numerical results.

15: \end{abstract}

16: