Collaborative optimization of production and energy performance in the coal blending management

A collaborative optimization solution is put up to integrate energy performance into the coal blending process for coking. Firstly, the collaborative optimization problem in coal blending management is described, and then the association model between the blending coal indicators vector and energy performance, production performance is provided by neural network ensemble technique to model their physical and chemical relation; the objective function and constraints of collaborative optimization model are derived from the association model. Secondly, the optimization model is figured out by genetic algorithm with the constraints expressed by non-fixed multi-stage mapping penalty function. Thirdly, the single factor sensitivity analysis procedure of energy performance is presented. The solution is verified through an iron and steel enterprises. The founded association model demonstrated the association relationships; Energy performance was optimized when the production performance is met, and more sensitive and less sensitive factors in the quality indicators of blending coal are achieved by the sensitivity analysis procedure.