The Decision Support System For Production And Logistics Operation Management In Metallurgical Enterprises

Metallurgical enterprise belongs to a kind of typical process industry, including the iron-steel, aluminum, bronze manufacturing and processing enterprise. In enterprise, the manufacturing and logistics operation management plays a very important role in decreasing cost, improving the quality of products, also in respond quickly to the market, and the goal is to minimize the total logistics cost and optimize the operation process. So, in order to reduce stock, decrease inventory cost, logistics cost, and increase customers satisfaction, under the background of the iron-steel enterprise, this article studies the ore-yard dynamic input location problem, meanwhile develops the 3D ore-yard simulation system, under the background of aluminum enterprise, we study the optimization problem with batch decision, as well as develop a decision support system (DSS). The main work of the thesis includes:(1) For the ore-yard dynamic input location problem, according to the characters of ore yard input location problem, we formulate it as an effective and dynamic MILP mathematical model, and solve the problem for the optimization by ILOG CPLEX software. While considering the practical scale, if only formulating one huge and unsolvable MILP problem, even through using the most advanced optimization software, it will consume enormous resource of computer and time for a feasible solution. Therefore, we decompose it into two small problems by integer and continuous variables so as to make use of the proven optimization algorithm Benders Decomposition to solve it. When the scale of the problem expands, the result is not very ideal. It is very important to rapidly improve the LB and UB that the speed solving main problem, the main problem termination rules, as well as how to generate strong effective cut. Based on this conclusion, we propose using intelligent optimization algorithm for solving the main problem, changing the termination rules to accelerate the speed, the test shows that the average gap between the suboptimal solution got by improved algorithm and the optimal solution is 0.045%, time in solution is greatly reduced.(2) Using the OpenGL technology for simulating 3D dynamic ore-yard, it can be realized through a control panel that rotating the stock within 360 angles, changing the elevation, the light position attribute and etc. It is easier for the management personnel to real-time access the ore-yard information in the remote. Therefore, it can improve the ore-yard operation information as well as modern management.(3) According to the characteristics of aluminum casting, formulate a mixed-integer linear programming (MILP) model for the scheduling of a multistage process using heterogeneous parallel machine. When solving the model with only 10 orders, it costs almost an hour to get a feasible solution. To solve the problem that the speed is too slow, the original problem is split into two sub-problems in a natural way using the special features of ore-yard. One is batch decision-making subproblem, the other is batch scheduling subproblem. Meanwhile we put forward a hybrid Tabu Search algorithm based on MILP, which contains of batch decision-making, heuristic and Tabu Search algorithms. The initial solution is obtained by batch decision-making algorithm combined with the heuristic algorithm, and it is improved intelligently by the Tabu Search algorithm. We construct Shift, Swap and 2-OPT neighborhood search strategies, and then improve the solution through variable neighborhood search strategy. The results show that it is proven that this hybrid algorithm can rapidly obtain good feasible solutions to solve the large-scale problem.(4) Under the background of a domestic aluminum production enterprise, we develop a DSS, through analysing the bottleneck of the production process, the embedded algorithm can give the production plan automatically and rapidly. The planning result provided by algorithm will be displayed in the interface of artificial scheduling system in the form of Gantt charts, which can be used directly by schedulers or can be taken as a basis of further adjustment.