Research On Less-than-Truckload Transportation Problem In Coordination And Its Solution Algorithm

In recent years,with the rapid development of global trade and e-commerce industries,national economy is increasingly affected by the transport industry.Less-than-Truckload transportation is a mode of transport for the delivery of scattered goods.Compared to the Truckload transport,the goods for each owner it needs to transport is less than one car.It can make full use of enterprise resources,efficiently fulfill customers' tasks and increase the efficiency of transportation companies.Under the increasingly fierce competition of Less-thanTruckload enterprises,in order to be more competitive,companies must satisfy the highquality shipper tasks while minimizing shipping costs.Based on the research work of many other scholars,this thesis studies the problem of collection and distribution transport for a variety of goods,mainly including the following contents:(1)Through reviewing a number of related domestic and foreign research literature,this thesis summarizes the Less-than-Truckload problems,vehicle routing problems and related algorithms.Firstly,the problem of hub-and-spoke Less-than-Truckload transportation and ordinary Less-than-Truckload transportation are mainly reviewed.Secondly,the solution algorithm is reviewed and introduced.The theoretical basis of the column generation algorithm and the framework of the simulated annealing algorithm are introduced in detail.(2)Based on a mathematical model aimed at minimizing the total transportation cost of a cooperative Less-than-Truckload problem in the reference,a column generation algorithm is designed.Firstly,the original model is decomposed into a main problem model and a subproblem model using Dantzig-Wolfe decomposition.Secondly,the sub-problem model is optimized,and an effective algorithm of column generation is designed.The experimental results show that,compared with the direct use of CPLEX software,the exact column generation algorithm can solve larger scale examples and solve the same-scale example more quickly.(3)According to the characteristics of the problem a solution algorithm is designed based on Threshold-Accept algorithm.The algorithm uses the interpolation method to generate the initial solution.The encoding method and two kinds of neighborhoods are designed.And according to the solution in the iterative process,temperature T is adjusted,preventing the algorithm from falling into local optimization.In the small-scale examples,the comparison between Threshold-Accept and column generation algorithm is carried on,and it proves the fastness of Threshold-Accept algorithm.In large-scale examples,the solving results of CPLEX and Threshold-Accept algorithm are compared and analyzed.It is obvious that ThresholdAccept algorithm can solve the problem in shorter time for large-scale cases and reduce the transportation cost by 20%.