Multi-level Evolutionary Genetic Algorithm for Solving VRPSPD Problem

In the field of industrial production and distribution, in order to meet the production needs of production enterprises, it is necessary to consider the customer’s need for simultaneous picking and delivery in the classic vehicle routing optimization problem, which we called VRPSPD(vehicle routing problem with simultaneous pick-up and delivery). In order to minimize the transportation cost and rationally use the distribution vehicle to balance the workload, establishing a VRPSPD target model that considers both the transportation cost and minimizing the maximum transport length difference between distribution lines, and constructing an improved genetic algorithm(GA) based on the characteristics of the model. The algorithm use the RMPCIM insertion method to improve the global search ability and accelerate the iterative process. At the same time, a new multi-level evolution iterative process is adopted in the population iteration to improve the superiority of the algorithm. The accuracy of the algorithm is verifified by the simulation of 18 sets of Solomon exextended instances. Compared with different algorithms in the literature,the result shows that the algorithm can weigh the objective function and obtain the optimal objective function.