基于KL-DE算法的抛物型参数识别

The Parameter Identification of Parabolic Equations Based on KL-DE Algorithm

为了降低将抛物型参数的维数,把参数假定为协方差已知的随机函数,并进行Karhunen-Loeve法分解,然后利用遗传算法求解此反演问题。通过仿真结果表明,改进算法可为偏微分方程的参数识别提供一个有效的工具。

In order to reduce dimension in the patameter inversion problems of parabolic equations,the patameter are assumed to be normal random functions with known covariance and decomposed by Karhunen-Loeve expansion,then obtain solutions by an differential evolution computational method.The simulated results show that the improved algorithm could provide an effective tool for the parameter identification of partial differential equations.