Research On Integrated Plan Of Railway Consolidation And Stowage Of Steel Product Based On Approximate Dynamic Programming

The coil transportation of railway includes two operations,i.e.,a consolidation plan and a stowage plan.The consolidation plan is to determine the plan of transportation according to the information of wagon and steel products in the warehouse,and stowageplan is to determine the specific loading position of the coil on the train.At present,consolidation plan and the stowage plan are made independently,which affects the loading efficiency of wagon because of the lack of systematic decision making.Therefore,based on actual process of transportation,this thesis raises an integrated method to solve the problem,studying the model and solving methods to improve the utilization of wagon and reduce the logistics cost.Based on the assembly logistics process of coils in steel enterprises,this paper investigates the integration problem of consolidation plan and stowage plan.The problem is to determine the specific loading location for each planned coils that collect from warehouse and an integer programming model is established for this problem and two kinds of dynamic programming algorithms that are proposed to.Then an approximate dynamic programming algorithm is proposed.The specific research contents are summarized as follows:(1)In order to solve the integration problem of consolidation plan and stowage plan,a linear integer programming model is established to maximize the total loading rate of wagons to determine that which coils need to be loaded and where the coil can be putted.As wagon can be loaded with different number of coils,there are different modes to place the coils.A position-based encoding strategy is designed to precisely express different modes of placing the coils on the trains.Based on the phases of loading one coil and one wagon,the established integer model is restructured to be two time sequence-based models.The standard optimization software CPLEX was used to test the random data and the actual data to verify the effectiveness of the models.(2)In order to solve the problem that CPLEX cannot solve large-scale instances within a reasonable time,this paper designs an approximate dynamic programming algorithm to solve the problem.According to the characteristic of multi-stage decisionmaking of the integrated project,the dynamic programming algorithm is designed from the perspective of the dynamic performance and the conversion from space to time.the approximate dynamic programming algorithm was designed based on the dynamic programming algorithm.The approximation strategy is designed and a single-stage submodel of integrated plan is established.And the smoothing strategy is used to update the approximate slope parameter used in the next iteration.Finally,based on the actual data and the random data.The experimental results show that the approximate dynamic programming algorithm can effectively solve the dimensionality problem of the dynamic programming algorithm and the solution time of solving problems is obviously shortened while solving large-scale problems,which proves the effectiveness and superiority of the algorithm.(3)Based on the established mathematics model of integrated planning problem and the designed approximate dynamic programming algorithm,an integrated planning decision support system for railway transportation is developed.The system includes these functions,providing the plan of program,manual adjustments,illegal inspection,evaluation of results and printing stowage results.The development of the system increases the actual collection and distribution efficiency of enterprise,and it fundamentally reduces the logistics and transportation costs of the enterprise.