The Conditional Markov Chain Search (CMCS) is a framework for designing metaheuristics for discrete combinatorial optimization problems. It decides the order of application of algorithmic components like hill climbers and mutations. However, CMCS lacks an acceptance criterion, making it good at exploration but not exploitation. This study explores extensions of the framework to improve its exploitation capabilities. The framework was applied to the three-index assignment problem, with results showing a two-stage CMCS superior to a single-stage one.
Publication date: 30 Jan 2024
Project Page: https://arxiv.org/abs/2402.00076v1
Paper: https://arxiv.org/pdf/2402.00076
