Research Of Multi-objective Evolutionary Algorithm Applications On The Vehicle Routing Problem In Logistics

With the rapid development of market economy, modern logistics distribution caused by more and more people’s attention. In order to meet the market’s needs, we should optimize logistics distribution system scientifically and reasonably. Vehicle routing problem is an important part of logistics distribution system. How to choose the vehicle distribution lines scientifically, efficiently and reasonably is an important approach to reduce cost and improve business efficiently for enterprises. Meanwhile, vehicle routing problem belongs to a NP problem of the combinatorial optimization problems. This problem has been a hot topic which the domestic and foreign scholars concern.In this paper we establish and study the two multi-objective optimization models of vehicle routing problem in logistics. Then it uses multi-objective evolutionary algorithm to solve the two problems respectively. And it gets the pareto-optimal solutions of the model eventually.Firstly, we give the review of the past researches on vehicle routing problems, we introduce the concept of vehicle routing problems, classification and solution algorithm. The paper introduces the vehicle routing problem with time windows, summarizes the optimization principle of multi-objective optimization problems. Then it expounds the differential evolution algorithm and the NSGA-II algorithm. The main research work of this paper is described as follows:(1) we establish a mathematical model of the unidirectional logistics distribution vehicle routing problem with no time windows. This is a multi-objective optimization model which contains the vehicle capacity restriction, the longest distance restriction. The solution must ensure the non-full load factor is least, and the total distance is the shortest. This paper uses an appropriate encoding method, it presents differential evolution algorithm for this problem, gets a series of pareto-optimal solutions. The article uses MATLAB language to simulate with computer, the computational results demonstrate that the differential evolution algorithm is effective.(2) we establish a multi-objective model to optimize VRPTW. This model contains the vehicle capacity restriction, the longest distance restriction and the least waiting and delay time of delivery vehicle. Compared to the model in the previous chapter, it is more practical significance. Meanwhile, we use the multi-objective evolutionary algorithm NSGA-Ⅱ to solve this problem. The advantage of the proposed algorithm is proved by the simulation based on MATLAB language, the computational results demonstrate that the differential evolution algorithm is effective.