SDVRP model for emergency logistics and evolutionary heuristic approach

Inspired by the idea of split delivery vehicle routing problem (SDVRP), this paper proposes a multi-objective, multi-period emergency logistics model, which minimizes the unmet demand, total delivery time and unbalanced supply among demanders. The weighted aggression is taken to compromise the three objectives. The chromosome code structure is designed so as to translate the delivery plans into gene sequences. Based on the feasible initial population generated by the roulette wheel selection, disaster-affected sites are clustered and the evolutionary optimization is performed by the procedures of selection, crossover and mutation. A numeric example is presented and the optimal solution is obtained to validate the model and algorithm.