Required Coke Quality Influence on a Coking Coal Mix Price Research. Linearization of a Coking Coal Procurement Optimization Model

摘要

The article performs the problem of simplification of a previously developed coking coal mix procurement optimization model. The model was developed at PJSC MMK (Magnitogorsk Iron and Steel Works), one of the leading Russian metallurgical plants, for improving the cooperation between technological and economical departments. It includes their background and can be used as an expert system for the decision making process. However, interconnections between coal quality indicators and coke quality indicators resulted in nonlinear mathematical program task. Therefore, the weak point of the model was the solution algorithm. Linearization method was used to fight the problem, which brought the task to linear program problem. Linearizing the problem allowed to employ the sensitivity analysis so that it became possible to answer several questions essential for the procurement department. What is the inclusion price for the supplier? How much does the coke quality improvement cost? Which suppliers are essential, which are surplus, which are lacking and so on. The simplex method used for solving the task improved the solution speed from twenty minutes to seconds and guarantied the global solution for the problem.