Scheduling Optimization System For Cold-rolling Continuous Annealing Line

Cold-rolling is the major process of steel production. Products of cold-rolling can directly enter the market as commodity, and also can be used as raw material of the tinning and chromium plating process. Since cold-rolling products can gain a high added value, the optimization of its production process is very important.The production scheduling problem is based on the background of continuous annealing line. Due to the complicated production process in iron and steel industry, the scheduling problem has been focused on by many researchers. However, continuous annealing line has been widely used for only a few years, whose corresponding scheduling problem has some new features. First, continuous annealing line integrates the individual production operations into one production unit which keep the annealing treatment from being interrupted. This continuous production way makes it difficult to decide the initial sequence, since it should satisfy the requirements of every operation. Second, there are many constraints that should be considered in the scheduling, such as annealing temperature, coil's width, coil's thickness and due date, etc. At last, more than one factor should be considered for evaluating the coil scheduling, including setup cost and tardiness time. Therefore, the problem is defined as a multi-objective problem.The continuous annealing scheduling problem mentioned above can be classified to be family scheduling problem. In this problem, what we should do is to decide the sequence of families and the sequence of coils within a family, since family has been predetermined before scheduling. A heuristic algorithm is proposed to solve this problem, and a worst-case analysis is also given to verify the efficiency of the algorithm. For small-scale problems, the solution obtained by heuristic algorithm is compared with the optimal solution obtained by standard optimization software. For large-scale problems, the heuristic algorithm is compared with tabu search algorithm, since standard optimization software can hardly find optimal solution in reasonable time.As some constraints have not been taken into account in family scheduling problem, a practical coil scheduling problem is considered here. A multi-objective mathematical model is established which takes into account all the production constraints, and it takes setup costs and total tardiness as objective function. The special case of this problem has been proved to be NP-hard. As a result, this problem must be a NP hard problem. For NP hard problems, exact algorithms and mathematical programming methods can not obtain the optimal solution in polynomial time. Therefore, we propose tabu search algorithm which has been successfully used in solving large-scale problem with complicated constraints. At the same time, the design of effective neighborhood improves the algorithm speed as well as the solution quality. Experimental results prove the efficiency of tabu search algorithm.Finally, we develop the scheduling optimization system for continuous annealing line. The tabu search algorithm mentioned above system has been embedded in the system to implement the scheduling function. With the help of data management function, the system has the ability to solve all the coil scheduling problems in continuous annealing process.