Research Of Production And Logistics Scheduling Problem In Steel Industry Based On Benders Decomposition

Benders decomposition is an effective algorithm to solve large scale mixed integer programming problems by decomposing a difficult mixed integer programming model to a master problem and a sub-problem. Based on the successful case for solving many practical problems, this paper focuses on the improvement strategies of Benders decomposition to the practical production and scheduling problems including a steelmaking continuous-casting production scheduling problem and a coil shuffling logistics scheduling problem with consideration of their special characteristics.The major works in this thesis are given as follows:1) Steelmaking continuous-casting process is one of the main processes in the iron and steel production, which is an integrated work of steelmaking, refining, continuous-casting. The steelmaking continuous-casting production scheduling problem is to determine the assignment of casting and heating to the machines, the schedule and the time table of the jobs on each machine. For the problem, the paper formulates a mixed integer programming model.2) Since this steelmaking continuous-casting production scheduling model has both discrete and continuous variables, the paper develops Benders decomposition algorithm to solve the problem. For improving the algorithm, the paper proposes heuristic algorithm to improve the initial solutions of the master problem and combinational Benders cut. The numeric results of the experiments based on lots of data show that the algorithm has good performance. In addition, a decision support system was developed to complete the scheduling work.3) The acid rolling warehouse holds the hot rolled coils. Shuffling is a process of removing the coils which block coils in demand which consumes lots of operation costs and time. The coil shuffling logistics scheduling problem is to determine the coil shuffling plan based on the production schedule. With consideration of its characteristics, a mixed integer programming model is set up.4) Since the coil shuffling logistics scheduling model involves both discrete and continuous variables, this paper adopts Benders decomposition to solve it, and develops the combinational cut to improve this algorithm. The experiment results based on practical production data verify the feasibility and effectiveness of the algorithm. Finally, a decision support system is developed which inserts the proposed model and algorithm into it.