多目标约束遗传算法在地质样品相态分析中的应用  

Application of Multi-target Restriction Genetic Algorithms in Phase Analysis of Geological Samples

作　　者：王洪艳[1] 杨兵[1] 

机构地区：[1]吉林大学朝阳校区化学系,吉林长春130026

出　　处：《岩矿测试》2002年第1期19-23,28,共6页Rock and Mineral Analysis

基　　金：国家自然科学基金资助项目 (2 9975 0 0 4)

摘　　要：多目标约束遗传算法 (MGA)用于地质样品相态分析 ,建立了相态浓度、浸取率常数及各相态浸取总量的数学定量关系。选择出铁的 5种相态的浸取剂和最佳浸取条件并确定出浸取率常数。讨论了多目标约束遗传算法的参数对计算速度及相态分析结果的影响 ,验证了方法的准确度 (|RE|<1 3 .75 % )和精密度 (各相平均RSD =4% ) ,建立了一种简便、快速、准确的地质样品相态分析的方法。In this article, Multi target Restriction Genetic Algorithms (MRGA) is introduced to phase analysis of geological samples. A mathematical model which reveals the relation among phase content, dissolved ratio constant and total dissolution content is founded. The optimal dissolving conditions for all phases are established. The effects of MRGA parameters on calculation speed and the analytical results are discussed. According to experiments on accuracy and precision for national standard samples (China), the average value of relative standard deviation (RSD) of all phases is 4% and the absolute relative error (RE) is less than 13.75% except for the phases with very low content. It proves that this phase analysis method for geological samples is simple, quick, accurate and practical.

关 键 词：多目标约束遗传算法 相态分析 地质样品 相态浓度 浸取率常数 

分 类 号：P575[天文地球—矿物学]