Model and algorithms for dynamic and stochastic vehicle routing problem

Abstract

This paper researched a dynamic vehicle routing problem of stochastic requests, in which at the scheduling moment t whether customers need service can be identified, but the demand is a random variable and submits to Poisson distribution. Based on model of M.H Lars, considering the multi-depot and open characteristics, two-stage stochastic programming models with resource are established. Then an adaptive immune quantum-inspired evolutionary algorithm (AIQEA) for this dynamic problem was proposed. In the AIQEA, in order to avoid the search process into the local minimum value an immune operator is introduced to optimise sub-routes. Finally, simulation examples are tested, and analyse the optimisation results of different values of the parameters l, experiment results show that this method can effectively solve the dynamic vehicle routing problem of stochastic requests.