改进的混沌遗传算法及其在炼铜转炉操作优化中的应用

Modified chaos genetic algorithm and its application on operation-optimization of copper smelting converter

利用混沌优化的遍历性和遗传算法优化的反演性 ,提出了混沌遗传算法 (CGA) ,其基本思想是把混沌变量加载于遗传算法的变量群体中 ,利用混沌变量对子代群体进行微小扰动 ,随着搜索过程的进行逐渐调整扰动幅度。结果表明 ,该方法优化效果与前人的优化结果相比 ,优化效率明显提高。由炼铜转炉造渣期与造铜期操作参数 (样本采集、数据预处理、PLS(偏最小二乘法 )空间变换、BPN神经网络建模及CGA)的优化和造渣期适应度函数与造铜期适应度函数的变换 ,使操作参数变量在训练集给出的数据范围的基础上延伸± 10 % ,得到最优点对应的工艺条件 ,并用于生产中。经过 3个多月的试运行 ,粗铜产量提高 6 .0 % ,冷料处理量提高 8% ,平均炉寿从原来的 2 13炉提高到 2

By use of the ergodic property of chaos movement and the inversion of genetic algorithm, a chaos genetic algorithm(CGA) was proposed. Its basic principle lies in the small disturbance of which extent is adjusted during searching to child generation group using the chaos variable. The results indicate that the CGA has good performance and significantly improve the computational efficiency in optimization compared with others. The optimization steps are as follows in slag making period and copper making period of copper smelting converter: sample collection, data pretreatment, space transformation using PLS(Partial Least Squares), modeling using BPN and optimization using CGA. The function of degree of adaptability is f(T)=(-|T-1?250|) max  in slag making period and the function of degree of adaptability is f(T)=(-|T-1?180|) max  in copper making period. The operation parameter variables are changed within the limits of ±10% in the scope of training samples. The copper productivity was improved by 6%, the mass of cold input was increased by 8% and the average converter life span was improved from 213 to 235. The economic profit reaches 26.4 million RMB produced by the increasing productivity of coarse copper.