Parallel Cooperative Memetic Co-evolution for VRPTW

Tackling complex transportation problems is a core issue in intelligent transportation systems, logistics, and planning. Such discrete problems commonly deal with multiple transportation goals mapping them into the objectives targeted by the optimization algorithm. We introduce the parallel co-evolutionary memetic algorithm for the vehicle routing problem with time windows, being one of the most widely-adopted rich vehicle routing problems. Our approach benefits from the parallel evolution of several subpopulations, each intensifying a specific optimization aspect of the local search phase in the underlying memetic algorithm, and from the process of migrating the best individuals across subpopulations. The extensive computational experiments indicate that our technique outperforms a parallel memetic algorithm in which the parallel populations use the same local search procedures.