Research And Application Of Production Scheduling And Its Optimal Algorithms On Steel Rolling

As one of the deep manufacturing procedures in steel industry, rolling process is a very significant working stage, which can increase the additional value of steel product, improve the entire competition power of steel enterprise and gain the related economic benefits for manufacturer. With the development and application of Computer Integrated Manufacturing System (CIMS) in steel industry, the running ability of Manufacturing Executive System (MES), based on Process Control System (PCS), makes a direct influence on production cost and competition power of enterprise. As for the main function of MES in steel industry, production planning and scheduling is playing a very important role on connecting production management with manufacturing job. Based on a key project of National High-Tech Research and Development Programme, this dissertation studies the production planning and scheduling problem of rolling process in detail taking into account the current production management mode of steel plant. The dissertation has mainly carried on the following research.A two-stage production planning and scheduling method of hot rolling is proposed in this dissertation based on the modern rolling manufacturing mode. Firstly, a production planning model for hot rolling batch plan is built up. The model is formulated as a class of Vehicle Routing Problem with Time Windows (VRPTW) under a couple of constraints, which is solved by a new PDPSO algorithm combined with the heuristic rules. Secondly, taking the planning results of the first stage as the input data, several batches of rolling plan are optimized by multi-intelligent hybrid algorithms.Based on the obtained scheduling solution of hot rolling, a charging planning method of walking beam furnace that works in front of hot rolling process is presented. This dissertation summarizes the charging plan as an earliness/tardiness scheduling problem, and solves it using a heuristic based evolutionary algorithm. Then, the real manufacturing data is used to verify the model and its related algorithm.The multi-type orders grouping to cold rolling process and its due date optimized method are studied. An order grouping scheduling model with fuzzy process time of machines and fuzzy due date is constructed due to the working time uncertainty of producing a group of orders. This model is summed up as a specific fuzzy Job Shop scheduling problem, in which the processing time of a group of orders is described as a trigonal fuzzy value that combined with the fuzzy due date function of this group to form the objective fuction of the model for global optimization.The cold batch rolling scheduling problem is studed in this dissertation. It is divided into two parts, coil merging and rolling planning. Firstly, the coil merging optimization, which is worked at the end rolling period, is carried on. This problem is formed as a Multiple Container Packing Problem (MCPP) and is solved by a new intelligent search algorithm—Discrete Differential Evolution (DDE). This algorithm has a relatively high computational speed guaranteeing the solution quality. Secondly, a Double Traveling Salesman Problem (DTSP) without priority model that separates a batch of plan into several rolling segments is used to solve the cold rolling planning problem after coil merging optimization.Based on the research mentioned above, a scheduling software system for rolling process in steel industry is developed. The application in Shanghai Baosteel Co. Ltd. shows that the system can improve the decision ability of production planning and scheduling in steel enterprise, decrease the production cost, save energy and reduce the working complexity of scheduling worker.