Vehicle Routing Problem Based On An Genetic Algorithm

With the continuous development of logistics industry and the constantly improvement of the specialization level of logistics in our country, logistics distribution industry has also been developed quickly. Of all the activities in distribution, the vehicle routing problem is the key problem. The windows is important in the optimization about the vehicle routing problem.The object of this paper is to research the single-depot vehicle routing problem with not full loaded. Firstly it introduces city logistics, then analyses the characteristics and the attributes of it and the major used areas of the vehicle routing problem. Above the basic vehicle routing problem, this paper founds the math model of the vehicle routing problem with time-windows.This paper adopts improved genetic algorithm to solve the vehicle routing problem with time-windows. The genetic algorithm chooses the nature data coding method, the initial group adopted mixed method which generated partly by randomly and partly by generation, selection strategy adopted the method of combining roulette wheel selection and truncation selection, and the improved edge recombination crossover instead the commonly used PMX, OX and CX. It got a satisfactory solution according to the termination of the conditions.Finally, the algorithm realizaed by C++, verified with the test data in the Solomon C101 customers in 25 sets of data, and compared with different algorithm, fond that the computer time cost and results have maked improvement obviously. The results in this paper which based on genetic algorithm in the city logistics and distribution vehicle routing problem areas and the development of the logistics system has some reference value.