Local search move strategies within MOEA/D

Local search (LS) is at the cornerstone of many advanced heuristics for single-objective combinatorial optimization. In particular, the move strategy, allowing to iteratively explore neighboring solutions, is a key ingredient in the design of an efficient local search. Although LS has been the subject of some interesting investigations dedicated to multi-objective optimization, new research opportunities arise with respect to novel multi-objective search paradigms. In particular, the successful MOEA/D algorithm is a decomposition-based framework which has been intensively applied to continuous problems. However, only scarce studies exist in the combinatorial case. In this paper, we are interested in the design of cooperative scalarizing local search approaches for decomposition-based multi-objective combinatorial optimization. For this purpose, we elaborate multiple move strategies taking part in the MOEA/D replacement flow. We there-by provide some preliminary results eliciting the impact of these strategy of the final population and more importantly on the anytime performance.