1: \begin{abstract}

2: We use sensitivity analysis to design \textit{optimality-based} discretization (cutting-plane) methods for the global optimization of nonconvex semi-infinite programs (SIPs).

3: We begin by formulating the optimal discretization of SIPs as a max-min problem and propose variants that are more computationally tractable.

4: We then use parametric sensitivity theory to design an efficient method for solving these max-min problems to local optimality and argue this yields valid discretizations without sacrificing global optimality guarantees.

5: Finally, we formulate optimality-based \textit{generalized} discretization of SIPs

6: as max-min problems and design efficient local optimization algorithms to solve them approximately.

7: Numerical experiments on test instances from the literature demonstrate that our new optimality-based discretization methods can significantly reduce the number of iterations for convergence relative to the classical feasibility-based method. \\[0.1in]

8: \keywords{Semi-infinite programming, Robust optimization, Discretization, Global optimization, Cutting-planes, Sensitivity analysis}

9: \end{abstract}

10: