Study On Heuristic Algorithms For Some Scheduling Problems Of Heat Treatment Process In Mill Roll Production

The mill roll industry is an upper-layer enterprise of the iron and steel manufacturing industry. The mill roll process flow is composed of the smelting, the molding, the casting, the heat treatment and the machining crafts and so on, in which the heat treatment plays a special important role because of the long processing cycle and high production costs in the mill roll process flow. In recent years, along with the rapid development of the iron and steel industry, the demand quantity of mill roll products and the production order quantity of the mill roll enterprise are increased continuously, which will lead to a reduction relatively for the heating quantity of the heat treatment. And then the heat treatment process becomes a bottleneck of the whole mill roll production flow. Therefore, under the condition of the existing resources, how to improve the utilization ratio of heating furnace in the heat treatment process, improve the production management level, and reduce production costs is very important to increase competitiveness of enterprises.In this dissertation, based on a production and information management development project of the machinery and mill roll enterprise in mainland China, and through investigating the actual situation of this enterprise and applying the relevant theories of the production management and the systems engineering, and then on the basis of establishing the system architecture of the mill roll production planning and scheduling, we establish the system function diagram of the mill roll production planning and scheduling. It lays a foundation for the following study on some optimization and scheduling problems of the mill roll heat treatment process. At the same time, it is a constituent part of the national ’985’project (2009CB320601) and the’National Natural Science Funds for Distinguished Young Scholar project’(70625001).To sum up, this dissertation focuses on the five key problems:(1) We discuss the process of the mill roll production planning and scheduling. On the basis of analyzing the characteristics of the mill roll production flow and some production management problems in the actual production, and applying the relevant theory of the production management, the system architecture of the mill roll production planning and scheduling is built. This system architecture includes five function modules:data management, production contract making, master production planning making, monthly production planning and job scheduling making.(2) We study a single-stage optimization scheduling problem with the rough mill roll batching and the batch scheduling for the heat treatment process. This problem can be seen as a parallel machines scheduling problem with batch loading. On this basis, we tried to combine three rough mill roll batching heuristics with three batch scheduling heuristics to solve the built mathematical model. In the simulation experiment, the effectiveness of the algorithm is demonstrated by the comparisons and analyses of results. At last, the optimization solving method is given for this problem.(3) We discuss a single-stage optimization scheduling problem with the batch assignment and the batch scheduling for the heat treatment process. This problem belongs to a parallel machines scheduling problem with assignment. With regard to the high temperature furnaces with different volumes, we present different assignment heuristics for both the parallel machines scheduling with specified assignment and the parallel machines scheduling with unspecified assignment respectively. We tried to combine all assignment heuristics with batch scheduling heuristics of (2) to solve the built mathematical model. The last simulation experiments show that the optimization scheduling results of the adjusting roller batches of shortest processing time with unspecified assignment heuristics are superior to those of other assignment heuristics.(4) We discuss a single-stage integrated decision-making problem with the rough mill roll batching, the batch assignment and the batch scheduling for the heat treatment process. In this problem, because the number of high temperature furnaces is equal to that of low temperature furnaces, we present the alternative batching assignment heuristics during the batch assignment. Furthermore, we develop a genetic algorithm with the sequence encoding during the batch scheduling. At last, the effectiveness of the algorithm is demonstrated by the comparisons and analyses of results, and some methods are given for the policy-maker to arrange production reasonably.(5) We discuss a multi-stage optimization scheduling problem with the same volume of heating furnaces for the heat treatment process. This problem can be seen as a no-wait hybrid flow shop scheduling problem. According to the no-wait characteristic between two sequential operations of the batch, the no-wait algorithm in stages is designed. Combined with the no-wait algorithm in stages, a discrete particle swarm optimization algorithm is developed to solve the integer-programming model. In order to assure the particle encoding still keep the sequential encoding we present a feasible solution adjustment method. At last, the effectiveness of the algorithm is demonstrated by the comparisons and analyses of results, and the equipment reformation strategies of actual reference value are given as well which is beneficial for the policy-maker to arrange production reasonably.