Vehicle routing problem with uncertain demand and travel time

To study stochastic Vehicle Routing Problems (VRP) with uncertain demand and travel time, a stochastic programming model was formulated and an improved genetic algorithm with self-adaptive mechanism was proposed for routes optimization. Self-adaptive selection mechanism was introduced for amending the fitness value to overcome the premature convergence effectively and a new crossover operator was developed in the algorithm. Computational simulations and comparisons based on two kinds of stochastic VRP problems with different size were provided. Results demonstrated that the proposed algorithm obtained better results with higher convergence speed.