A Memetic Approach for Routing Problem with Capacity and Time Constraints

In this paper, we propose a memetic approach to deal with routing problem with capacity and time constraints. Our approach aims to determine least-cost trips for a homogenous fleet of vehicles that serve a set of geographically dispersed customers with deterministic demands. Customers time constraints and vehicle capacity have to be rigorously respected during the solving process. Our approach is based on a bio-inspired stochastic algorithm that mimics the reaction process of chemical molecules, which interact until attaining a stable state with minimum free energy. Through the sequence of elementary reactions, molecules explore different regions of the solution space toward the lowest energy state. Since the effectiveness of the optimization process mainly depends on the quality of initial population, we integrated a two-step procedure involving greedy randomized construction with local search. In order to assess our approach, we present a wide variety of computational experiments. The obtained results are compared to those of the best metaheuristics. These results confirm the effectiveness and good performances of our approach.