Research On Hybrid Evolutionary Algorithm To Solve The Vehicle Routing Optimization Problem

With the development of socialism market economy, logistics being the third profit headspring has evident influence on economy activities, and the logistics distribution industry obtained the rapid development. Logistics vehicle optimization schedule is the key cache for logistics distribution, it affects mostly on raising service quality, reducing logistics cost and increasing economic benefits of logistics distribution enterprises. In daily life and production, many problems such as mail delivery problems, bus scheduling problems, electricity dispatching problems, pipeline laying problems, robot path planning and computer network designing problems may be regarded as logistics distribution vehicle scheduling problems in the abstract.As a NP-hard problem, the logistics distribution vehicle scheduling problem will increase exponentially along with the adding of customers. As a probabilistic search algorithm based on biology evolution mechanism, evolutionary algorithm has showed that it has well capability for solving complex system optimization problems, especially for combinatorial optimization. Therefore it becomes an important studying trend to solve the problem with evolutionary algorithm.A hybrid evolutionary algorithm (HEA) is presented in this dissertation, to solve the vehicle routing problem (VRP) with VRP model based on comparison internal and overseas present situation and technology, and an intelligent logistics distribution system was developed that based on the HEA.Firstly, the memory function and the immune cell that produced by the immunity algorithm (IA) as the operator were joined in the genetic algorithm(GA) to seek the multi-peak value, then the Pareto optimal solutions were combined to distribute evenly for initial population, the question of precocious were solved in the local by simulation annealing algorithm(SA), and the strong ability of routing optimization and the positive feedback of the ant colony optimization (ACO) were added to reduce searching time.Secondly, simulation experiment were carried separately in the small and medium scale customer of the fixed region, and the simulation environment of electronic map were established in the experiment, according to the requirement of concrete task, the trajectory which connect the initial point to the end point and avoid the obstacle were searched in the environment of different density and distribution complex degree, the approximate optimal feasible path can be searched.At last, intelligent logistics distribution system were designed and developed with HEA according to the object-oriented system analysis and the design method, and applied in the Yanji city Jilin Province. HEA were combined with GA,IA,ACO and Pareto optimal solutions, and reduced the incomplete convergent phenomenon, and avoided falling into precociously, and solved two shortcomings of tradition genetic algorithm, it is efficient versatility.The results of simulation experiment show that the HEA can gain higher global convergence rate and speed. The intelligent logistics distribution system based on HEA can complete and satisfy the delivery requirement of delivery center in time and effectively. The system has certain practical value.