Regularization Particle Inverse Algorithm of Dynamic Light Scattering Based on Genetic Algorithms

摘要

Tikhonov regularization is an effective way in particles size inversion of dynamic light scattering (DLS). Regularization parameter choice is the key in this approach. In order to obtain the optimal regularization parameter, according to the Morozov discrepancy principle, an regularization inversion method based on genetic algorithms (GA) is proposed. It optimizes the regularization parameter within the global region by genetic algorithms, does not require a priori particle size distribution (PSD) and doesn't also define the initial value of the regularization parameter. Computer simulations datum of particles were respectively inversed by this method. The results show that when the noise level is 0～0.001, results of unimodal distribution particles agree with the theory distribution, and double-peak feature of bimodal distribution particles is clear, maximum error of theirs inversion peak value is less than 6.67%. Therefore, regularization inversion method with GA technique has strong immunity of noises and is feasible in particles size inversion of DLS.