Logistics is an emerging discipline and distribution is an important element of modern logistics. Well arrangement for vehicle routes distribution can cut down transport cost and improve efficiency. Vehicle Routing Problem (VRP) is a combination optimization problem in transportation logistics, which has been applied in many fields. It is a strong NP problem. Pickup and Delivery Problem with Time Windows (PDPTW) is more difficult to solve than VRP with simple conditions.In this paper, Genetic Algorithms (GA) for VRP with simple conditions was introduced and cross operator in GA was improved. These researches made full preparation for us to study PDPTW.The mathematical model of Pickup and Delivery Problem with Time Windows was analyzed and a deep study of Grouping Genetic Algorithm (GGA) was made. A taboo thought was used in insertion heuristics to produce feasible solution. A fitness function was defined for multi-objective programming. A new data structure was designed in this paper to implement GGA.Path adjustment strategies were proposed in GGA, and Multi-Strategy Grouping Genetic Algorithm (MSGGA) was designed. A benchmark data set for PDPTW was used to test MSGGA. Compared with other algorithms the experimental results have shown that MSGGA could find better solutions and take less time for computing. This proves that MSGGA is steadier to reach the solution of the problem. |