摘 要：利用多波长消光法，在已知粒度分布前提下，采用遗传算法 对折射率进行了反演。相比于传统的优化方法通常获得的局部最优 解，该算法获得的折射率结果逼近全局最优解。反演采用的波长数 目和颗粒单分散性的优劣影响解的可靠性。尽量多地获得透射谱信 息以及增大颗粒系的分散性可以是测量结果更准确。由于亚微米和 纳米级颗粒的消光系数 7 折射率曲线的单值性关系比微米级颗粒 更为明显，导致其目标函数的多峰分布程度得以降低，使反演的抗 噪能力增强，折射率更易于获得。 

关键词：颗粒测量；89 散射；折射率；反演

Abstract：The genetic algorithm (GA) is employed to inverse the refractive index of a particulate system. The particle size distribution is provided to be known in advance. Compared with other traditional optimization methods, GA's results approach the global optimum solution instead of locals ones. The effect of GA is influenced a lot by the number of wavelength used for inversion as well as the dispersity of the particulate system. More information of the transmission spectrum is known and the better the particles are dispersed, the higher precision of the results can be achieved. Besides, since the monotonicity of extinction coefficient-refractive index curve for submicron level particles is more obvious than that for micrometer level ones, the extent of the multimodal distribution will be weakened and the inversion algorithm is more resistant to the measurement errors, which makes the inversion results more reliable.

Keywords：particle sizing;Mie scattering; refractive index;inversion