This paper discusses MATRS, a new matrix adaptation trust region strategy for solving noisy derivative-free mixed-integer optimization problems with simple bounds. MATRS repeatedly cycles through five phases, mutation, selection, recombination, trust-region, and mixed-integer in this order. But if in the mutation phase a new best point (point with lowest inexact function value among all evaluated points so far) is found, the selection, recombination, and trust region phases are skipped. Similarly, if the recombination phase finds a new best point, the trust region phase is skipped. The mixed-integer phase is always performed. To search for new best points, the mutation and recombination phases use extrapolation whereas the mixed-integer phase performs a mixed-integer line search along directions going into a valley. Numerical results on several collections of test problems show that MATRS is competitive with state-of-the-art derivative-free mixed-integer optimization solvers
Article
View Heuristic methods for noisy derivative-free bound-constrained mixed-integer optimization