Improving neighborhood exploration into MOEA/D framework to solve a bi‐objective routing problem

Abstract Local search (LS) algorithms are efficient metaheuristics to solve combinatorial problems. The performance of LS highly depends on the neighborhood exploration solutions. Many methods have been developed over years improve efficiency different problems operations research. In particular, strategy and exclusion irrelevant neighboring solutions design mechanisms that be carefully considered when tackling NP‐hard optimization An MOEA/D framework including an LS‐based mutation knowledge discovery is core algorithm used a bi‐objective vehicle routing problem with time windows (bVRPTW) where total traveling cost waiting drivers minimized. We enhance classical from literature scheduling propose new metrics based customer distances times reduce size. conduct deep analysis parameters give fine tuning adapted variants bVRPTW. Experiments show proposed strategies lead better both Solomon's Gehring Homberger's benchmarks.