The adaptive convexification algorithm for semi-infinite programming with arbitrary index sets (Q1925776)

A feasible point method for semi-infinite optimization problems with arbitrary, not necessarily box-shaped, index sets is presented. The authors generalize the ideas of the algorithm presented by \textit{C. Floudas} and \textit{O. Stein} [SIAM J. Optim. 18, No. 4, 1187--1208 (2007; Zbl 1216.90094)] to multidimensional index sets (\(Y\)) and introduce an additional concept to handle sets \(Y\) with arbitrary shape. As the solution concept is that of stationary points, no global assumptions are made on the structure of the objective function or the constraints. The techniques presented for handling the set \(Y\) differ from those discussed in [\textit{P. Lemonidis}, PhD Thesis, Massachusetts Institute of Technology (2007)] in the adaptive way of the used subdivision strategies. A numerical example illustrates the performance of the method. In the last section of the paper, some final remarks are given and some possible improvements are pointed out.