Research On Slab Designing And Matching Optimization Problems In Iron & Steel Enterprise

In recent years, because of the increasing globalization of competition in the steel industry, the customer requirements have been more and more diversification and present the characteristics of low-volume, high-variety, and complication. On the other side, due to the production technology and processing equipment constraints, mass production has gradually become one of mainstream production modes in the steel industry, whereas it clashes with the customer demands. In order to lower the influence of such conflict, it is necessary for planners to optimize slab designing and slab matching in the steel process. Therefore, these two problems have been investigated in this paper as follows.(1) Slab designing problem is to map multiple orders with the same or similar specifications (including physical and chemical properties) into one large slab so that the number of mapping orders is maximized, and the number of designed slabs as well as the waste is minimized. To solve the slab designing problem, two heuristic algorithms which are respectively based on clustering and packing have been proposed, and tested on the practical production data. Computational results show that the packing-based heuristic algorithm can yield better solutions.(2) Slab matching problem is to match customer orders with inventory slabs so as to minimize the production cost. The open-order slab matching problems existed in the literature focus on the single-line case, while our considered problem in this paper is based on multi-line. The main difference is that the size of the instance is relatively large, i.e., the number of both orders and slabs is too large comparing with the single-line open-order slab matching problem, and the matching in different lines is coupled with each other. Moreover, the object of this problem is the important orders (which are proposed to match slabs in priority by planners). And the slabs ready for matching can be both the open-order ones and others with orders on them. The multi-line slab and important orders matching problem is formulated as a multiple objective 0-1 integer programming model, which is to optimize the matching between slabs and orders, as well as improve the rate of utilizing slabs under the consideration of slab priority, order priority and order integrality.(3) Due to the large-scale and complexity, scatter search, a new intelligent optimization algorithm and the one embedded with iterated local search algorithm are developed to solve the problem. In this algorithm, multiple heuristic algorithms (including the know-how algorithm based on the matching process) are proposed to construct the initial population. Several high quality and diversified solutions are chosen to compose the reference set. Then two of the solutions existed in the reference set are selected to be combined into a new solution. The iterated local search with the swap and insert neighborhoods is then used to try to improve the combined solution. In this hybrid algorithm, the reference set is updated based on the static update strategy. The hybrid algorithm has been tested by using practical data set collected from the steel company. Computational experiments demonstrate that the hybrid algorithm can generate the satisfactory solutions comparing with scatter search and the Know-how heuristic algorithm.(4) The decision support system embedded with the formulation and the hybrid algorithm has been developed and implemented in a certain domestic iron & steel enterprise. This developed system can help to improve the matching qualities, enhance the slab utilization and reduce the surplus weight and cutting weight of slab, which consequently increases the economic and social efficiency for the enterprise.