Research on Open and Dynamic Vehicle Routing Problems Based on Joint Distribution
GE Xian-long,WANG Xu,DENG Lei
DOI: https://doi.org/10.3969/j.issn.1004-6062.2013.03.008
2013-01-01
Abstract:logistics distribution has experienced fundamental changes caused by the integration of regional economic,product manufacturing network,dynamic customer demand,and traditional mass mode.The batch mode of the logistics distribution cannot adapt to the current complex logistics distribution requirements.Multi-set distribution,interval cross distribution,and joint distribution are emerging as new logistics distribution problems. The paper introduces the concept of time-axis to transform these dynamic problems into a static problem,and establish an open vehicle routing model.This model considers vehicles’ on-load rate to research cross-region multiple distribution centers,multiple kinds of and open joint dynamic distribution problems.Joint distribution problems involve multiple distribution centers and traditional vehicle routing problems.Multiple distribution centers adopt the logistics distribution mode in which a distribution center is responsible for a district logistics demand;however,the one-to-many mechanism of segmentation distribution has made sharing difficult between different distribution centers,and can lead to vehicle idling,no-load return and circuitous transportation problems. This paper proposes the whole logistics distribution mode,in which form different distribution centers can dock,replenish and respond to customer demand form distribution networks,and goods can be removed at will or distributed between different distribution centers.This kind of to many mechanism of network distribution strengthens interregional coordination and resource sharing logistics,and would be a good solution to balance resources in logistics distribution.To solve multiple type problems of in joint distribution,this paper puts forward the optimize vehicle route before matching types of vehicles principle of multiple types.This paper further establishes an evaluation function of fuel consumption using the cargo rate to choose the best type of in different distribution routes,which better solves previous routing problems of multiple vehicle types. Due to the high demand for timeliness and stable algorithms to solve dynamic distribution model of the joint distribution,this paper adopts the cloud model x generator to produce dynamic crossover and mutation probability,improve the fixed set pattern of crossover and mutation in standard genetic algorithms,and enhance the global searching ability and the fast convergence of the algorithm.Combined with the problems of dynamic joint distribution established,this paper designs the two-staged strategy initial optimization and real-time optimization. Finally,the paper analyzes the model and algorithm using the random generation dynamic joint distribution network.The experimental results show that the two-staged solution algorithm can respond to dynamic demand timely and provide a good solution for dynamic joint distribution of vehicles.In the experiment Ⅱand Ⅲ,the cloud genetic algorithm,standard genetic algorithm and taboo search algorithm are compared and analyzed.The results show that the cloud genetic algorithm can achieve global search and fast convergence simultaneously.Thus,the proposed algorithms could help solve dynamic joint distribution problems.