Optimization And Application Of Coal Blending Model In Magang

There are more than ten kinds of coal needed in coking process,and the price of different kinds of coal is different.Linear programming model is usually used to solve the coal blending problem,but the traditional linear programming model only solves the coal blending problem theoretically.In practice,there will be some errors due to the production process,resulting in a certain gap between the actual value and the theoretical value.The traditional linear programming model is improved by the feedback of the effect of the error to the model,and a set of efficient coal blending system is developed to improve the production efficiency of the enterprise.The main innovations of this dissertation are as follows:(1)Error analysis is used to make error regression for actual production data and planned production data.The planned mixed coal data of Magang is calculated by the traditional linear programming model.There is a certain error between the calculated theoretical value and the value of production.In this dissertation,the error between the theoretical value and the actual value is regressed,which provides a new way for Magang to find more accurate quality coke ratio.(2)By using the method of dividing feasible domains,Hadoop is used to solve the problem that the traditional gradient descent algorithm is easy to fall into local optimum for solving linear programming problems.The feasible region of the linear programming coal blending model is partitioned into several sub feasible domains,and Hadoop is used to calculate the sub feasible region.Then the optimal solution of each sub feasible region is sorted to find the global optimal solution.(3)Taking the coal blending model as the core,based on the actual production situation of Magang,a set of expert system for coal blending is set up, which integrates market,procurement,inspection,logistics,inventory,coking and iron making.Through coal blending operation as the core of production chain,we can not only guide procurement volume and safety stock,but also guide coke output and iron and steel output.The system can realize real-time management and automatic updating of coal,coke and steel and other related information and it provides strong support for decision-making and scientific management.