Solving a Stochastic Multi-Objective Location-Routing Problem: a Comparative Study of Meta-heuristics

This paper presents a multi-objective location-routing problem with capacitated vehicles to reduce the total cost of the system. The model considers times of traveling, service and waiting by vehicles while guarantees the least probability that the sum of these parameters be lower than a predetermined value when minimization of this value is considered as an objective function. Due to NP-hardness of the model, we propose a multi-objective imperialist competitive algorithm (MOICA) to solve the given problem. The great efficiency of the proposed MOICA is demonstrated via comparing with two famous meta-heuristics, named Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) and Pareto archived evolution strategy (PAES). After adjustment several crossover and mutation strategies for each algorithm based on response surface methodology, the associated results in terms of four existing comparison metrics on several benchmark instances indicate that the proposed MOICA outperforms the others

Keywords:

Location-routing problem (LRP) , Facility location , Vehicle routing , imperialist competitive algorithm (ICA) , Linearization