Research On Production Planning And Dynamic Scheduling For The Whole Process Of Cold Rolling And Its Application

One of the key functions is the production planning and scheduling of management information system of cold rolling enterprise. The management level and the production capacity could be improved with rational and feasible planning. However, since the establishment of planning and scheduling schema is largely restricted by the capacity of mills, the inventory of work in process, the specification and quality of products and delivery date, the schemas have to be with the characteristics of dynamic, adaptive and robust. Depended on a key project of National High-Tech Research and Development Programme, this dissertation studies the planning and scheduling problem for cold rolling process in accordance with the current production and management mode. The main contents of this dissertation are as follows.The scheduling problem of cold rolling mill is summarized as an asymmetric double traveling salesman problem that is implemented by job scheduling of cold rolling process due to the production specification of mill. The multi-objective scheduling model based on the Pareto optimization is constructed which includes three objectives, i.e. the change in width, thick of entrance and exit between two continuous rolling steel coils. A Pareto Multi-Objective Ant Colony Algorithm (PAACO) depending on Pareto non-dominated set is then proposed. The final scheduling result is obtained from the non-dominated set.A Subspace Adaptive Clustering with Interval Feature on Time Window (SACIT) is presented for grouping the orders according to their specifications. Based on the connection among mills in BaoSteel Cold Rolled Sheet Mill, the tandem cold mill is regarded as the center of the enterprise and the batch planning method of the whole flow is presented which considers the capacity of cold rolling mill and process specification. Using the proposed Time-segment Ant Colony Optimization (TACO), the orders production planning and capacity allocation are carried out taking account to the delivery date, the inventory of work in process and the production route of products.The partly reconfigurable hybrid Petri-net model is established to reduce the interference by the periodic maintenance and regular technology adjustment of some mills. A Finite-Searching Ant Colony Optimization (FSACO) is introduced by means of mill capacity, delivery date and affectivity of orders. In the period of maintenance and regular technology adjustment of mills, the production flows of part of orders are reallocated via the method and the proposed FSACO searches orders with high priority and make them close to due date in the restriction of mill load.Through the researching on the characteristics of dynamic events in the process of production, the dynamic planning and scheduling system based on the multi-agent is established with hybrid Petri net and UML technology. The different dynamic rescheduling non-linear programming models are constructed to the whole flow production and the reconfigurable mills. A Dynamic Constraint Ant Colony Optimization (DCACO) and a Contract Fetch Approach Based on Ant Colony Clustering (CFAAC) are presented to obtain the dynamic scheduling result.The production planning and scheduling system for the BaoSteel Cold Rolling Sheet Mill is established based on the above-mentioned models and algorithms and is applied to the practical production process. The results indicate the presented method in this dissertation can improve the strategy capacity of the planning and scheduling system, reduce production cost and the complexity of scheduling process and enhance the adaptability of the system.