| With the development of globalization and informatization,there is a profound shift in the market conditions.Logistics,as "the third source of profits",is concerned by a growing number of enterprises after reduce consumption of raw materials and increasing labour productivity.Distribution is an important element in logistics system,in many cases,the transportation cost is often the largest part of logistics cost.Therefore,it is one of the hot issues to reduce logistics cost by optimizing vehicle route.In most of VRP research,a condition has always been defined that one customer must be serviced by one vehicle.However,in real operations,there will be the waste of transportation capability if a majority of customer have heavy demand.In this case,cost saving could be gained by spilting the customer demand.The Spilt Delivery Routing problem is a variant of the Capacitated Delivery Routing Problem.This problem relaxes the constraint that the demand of one customer can only be serviced by one vehicle.And in real operations,customer demand is often not single,so we choose the Spilt Vehicle Routing Problem with Simultaneous Pick-up and Deliveries as the topic of this paper.This paper focuses on several aspects of the problem of the Spilt Vehicle Routing Problem with Simultaneous Pick-up and Deliveries:Firstly,we find out the general situation of the research of the spilt vehicle routing problem,and overview the vehicle routing problem.They are the basic of the research in this paper.Secondly,overview the spilt vehicle routing problem,analyze the property of result and research significance of feasible solutions to the spilt vehicle problem.Based on the research and the real situation,we established a mathematical model about the problem.Thirdly,based on the mathematical model,we design an algorithm using ant colony optimization to solve this problem.In this algorithm,we design the corresponding state transition rule and the principles of spilt points.At last,taking a test example to prove its effectiveness and the experimental results on a benchmark set show that the algorithm of this paper is competitive with a state of the algorithm.Finally,we summarize the main achievement of this paper and point out the future research direction. |