烧结配料优化模型求解方法研究  

Research on Methods of Solution to Model of Sintering Burdening Optimization

作　　者：吕学伟[1] 白晨光[1] 邱贵宝[1] 黄青云[1] 何生平[1] 

机构地区：[1]重庆大学材料科学与工程学院,重庆400044

出　　处：《应用基础与工程科学学报》2008年第5期695-702,共8页Journal of Basic Science and Engineering

基　　金：国家自然科学基金项目(50504020);重庆大学研究生创新基金(200701Y1B0020188)

摘　　要：针对烧结配料优化模型的求解问题,研究了线性规划法、蒙特卡洛法和遗传算法各自的特点.以国内某钢铁企业的烧结工艺为例,通过改变约束条件、模型求解参数等对比、分析求解结果,分别从最优解精度、约束条件、应用范围以及求解效率4个方面进行比较.重点研究了遗传算法的惩罚函数,提出了通过改变惩罚函数大小实现按照优先级处理约束条件的新方法,指出了该方法在配料优化中的潜在应用价值.To solve the optimization-model of sintering burdening problem, three different methods, linear programming, Monte Carlo algorithm and genetic algorithm, were used in the same situation. Take a local iron and steel company as a example, the results were compared and analyzed through changing the constraint conditions and models parameters, so three solutions were analyzed from the accuracy of final results, the constraint conditions, the using field and the effciency et al. The characters of each method were concluded. In the article the penalty function of GA （genetic algorithm） was analyzed mostly, and then the new idea was put forward which is to manage the constraint conditions by the priority using changing the penalty function, the potential application value of GA in sintering burdening optimization was also give out.

关 键 词：烧结配料优化 线性规划法 蒙特卡洛法 遗传算法 

分 类 号：TF541[冶金工程—钢铁冶金]