品位组合优化的粒子群—神经计算方法  

A Method of Integrating PSO and ANN for Optimizing Cut-off Grade and Grade of Crude Ore

作　　者：贺勇[1] 诸克军[2] 胡承凡[3] 

机构地区：[1]广东工业大学管理学院,广州510520 [2]中国地质大学经济管理学院,武汉430074 [3]武汉钢铁集团矿业有限责任公司大冶铁矿,湖北黄石435006

出　　处：《系统管理学报》2010年第2期204-209,共6页Journal of Systems & Management

基　　金：国家自然科学基金资助项目(70573101);武钢科研课题(070429);2009年广东工业大学博士启动项目

摘　　要：根据矿山采选物理过程,建立以截止品位与入选品位为决策变量,经济效益为目标函数的非线性模型,并采用粒子群-神经集成的方法进行优化求解。其基本操作为：截止品位与入选品位组合构成进化计算的个体,个体包括两部分,前半部分代表截止品位,后半部分代表入选品位;用BP网络和RBF网络建立收益（适应度函数）与粒子个体的局部联系;利用粒子群算法的全局搜索功能找出使适应度函数最大时的品位组合（截止品位及入选品位）。研究表明：当前大冶铁矿截止品位18%,入选品位41%~42%的生产方案有待改进,当截止品位为17.833 7%~17.836 7%,入选品位取值为46.4%,2007-01~2007-11的精矿量增加13.92万t,总净现值增加6.698百万元。According to the mining milling production of metal mine,this paper establishes a nonlinear model,in which cut-off grade and grade of crude ore are decision-making variables,economic benefit is on（objective） function,and integrates PSO and ANN to optimize the Cut-off grade and Grade of crude ore.The（idea） is detailed as follows： put Cut-off grade and Grade of crude ore together as particles for evolution computation,a particle stands for a combination of cut-off grade and grade of crude ore;use BPNN and RBFNN to get the local connection between the income value（fitness function） and chromosome;utilize PSO globally to search the optimal Cut-off grade and Grade of crude ore,to make the fitness function get the most value.Take Daye Iron Mine as an example,the result shows that： During the period of January to November in the year 2007,the optimal Cutoff grade is 17.833 7%～17.836 7%,and optimal Grade of crude ore is 46.4%,the optimized scheme can increase the amount of concentrate by（139 200） ton,and improve the net present value by RMB 6.698 million.

关 键 词：品位优化 粒子群算法 神经网络 截止品位 入选品位 

分 类 号：N945.15[自然科学总论—系统科学]