Study On The Modeling And Optimization Method Of The Torpedo Car Scheduling In Hot Metal Transportation

The scheduling for molten iron logistic optimization is to determine the distribution of molten iron from ironmaking stage to steelmaking stage,the transportation routes of torpedo cars and locomotives such that the molten iron from blast furnaces can be delivered to the steelmaking shops on schedule and on quality.When making the schedule of molten iron logistic optimization,the following key elements including the tapping information of blast furnaces,the demand of converters,and the status of preprocess positions such as deslagging,desulfuration and dephosphorization,should be taken into consideration.As the hot metal transportation process involves a series of processes,such as iron tapping,heavy pot transportation,pouring and empty pot transportation,the routing path is restricted by fixed rails and so on,which make the problem difficult to optimize.So the study of torpedo scheduling problem is of great significance to the molten iron logistic optimization.By taking the molten iron transportation process of an iron and steel enterprise as research background,the dissertation investigates the task allocation of torpedo car scheduling problem and the route of torpedo car scheduling problem,and the effective models and algorithm are proposed.By embedding these,the decision support systems software is developed.The whole research includes the following aspects.(1)For the task allocation of torpedo car scheduling problem,taking the linear relationship between energy consumption of molten iron and waiting time into account,a mix integer programming model minimizing the total weighted completion(waiting)time is formulated.Then the Dantzig-Wolfe decomposition approach is used to reformulate the MIP into a set partitioning master problem and price subproblems.The column generation algorithm is designed for the linear relaxation of the MP.In the implementation of the column generation algorithm,the heuristic based on the nearest neighborhood idea is conducted to get the initial feasible solution.According to properties of subproblems,the dynamic programming algorithm based on label correcting is constructed to solve subproblems.The branch and cut algorithm is developed and the branch strategy of it is based on the decision variable of MIP in fractional solutions to get the optimal integer solution.The column generation algorithm has been tested using the practical production data collected from an iron&steel complex in China,the computational experiments are carried out and the computational results verify the efficiency of algorithm.(2)For the marched route of torpedo car scheduling problem,after analyzing the traveling track constraints,the difficult issue of this problem is to deal with collision conflict in the process of transportation.According to the time-space characters of torpedo car scheduling,a time-space network flow model is developed.Then model is solved and the computation result indicates that the proposed model can effectively solve the torpedo car scheduling problem.(3)By embedding the models and solution approach,the decision support systems software is developed.The main functions of the system are proposed according to the demand analysis of the scheduling system for optimizing the logistic the molten iron.Meanwhile,the scheduling plan is shown by Gantt chart.