遗传算法在反演颗粒粒径分布中的应用  

Application of genetic algorithm in inversing particle size distribution

作　　者：马静静[1] 蒋小平[1] 彭铃舒[1] 

机构地区：[1]中国矿业大学机电与信息工程学院,北京100083

出　　处：《计算机工程与设计》2011年第3期1051-1054,共4页Computer Engineering and Design

摘　　要：为了提高颗粒粒径测量的抗噪性,提出了一种基于遗传算法的颗粒粒径分布反演问题的解决方法。基于Mie散射理论,建立了颗粒粒径测量模型,用所设粒径分布对应的光强分布与实测光强的差值作为目标函数,采用非均匀变异算子提高了算法的局部搜索能力。对数学模型进行Matlab编程,利用遗传算法使目标函数达到极小值,然后得到对应粒径范围的概率,从而得以确定颗粒粒径分布。数值模拟仿真结果表明,相比传统的反演算法,该算法具有更好的抗噪性。To improve anti-noise in particle size measurement,an inverse solution for particle size distribution inverse problem based on genetic algorithm is presented.Based on Mie scattering theory,a particle size measurement model is formulated.Fitness function is the difference between the intensity distribution corresponding to the given particle size distribution and what is actual measured,and by using nonsymmetrical mutation method the local searching ability of algorithm is improved.Firstly,Matlab programme for mathe-matical model is implemented.The genetic algorithm（GA） is used to make fitness function reach minimum.Then,probability in each particle size range is gotten.So the particle size distribution is determined.Results of the simulation indicated,the proposed algorithm has better anti-noise than traditional inverse algorithm.

关 键 词：抗噪性 粒径测量 粒径分布 反演问题 遗传算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]