南方煤矿精确选煤配售系统建模的研究

Research of the Accurate Placement System Modeling for the South Coal Mine

南方煤矿矿区产煤具有量少、种类多且煤质差异大的特点,在客户需求量大时,选择合理的煤炭配售方法建模是非常必要的。基于遗传算法设计的配售系统通常是在给定多约束线性规划的模型下求最优配煤比;而对模型参数的选取欠缺考虑。在已知配煤目标的前提下,不仅给出最终的决策方案,并且利用贪心算法与Powell搜索法结合的思想自动筛选出合适的单煤信息,作为精确配煤的参数。实验结果验证了系统的可行性和有效性。该系统已在福建煤电股份有限公司得到了成功的应用。

It＇s quite necessary to choose an appropriate method of coal placement modeling in the south coal mine,whose products have the feature of less product,different kinds and big diversity in quality.The placement system designed by genetic algorithm always obtains the optimal solution to the given model of the multi-constraint linear programming directly,but considers less about selecting the model parameters.The optimal solution ant t the appropriate single-coal information as accurate parameters were worked out by the combination of greedy algorithm and Powell search method under known coal blending targets.The result of an experiment has improved validity and feasibility of this system.The system has been used in Fujian Coal ＆ Electricity Group Co.,LTD successfully.