利用改进微分进化算法实现线性系统逼近

A modified differential evolution algorithm for linear system approximation

提出一种基于改进的微分进化算法的逼近算法。新算法通过参考粒子群算法惯性权重思想,引入惯性加权系数,在计算初期能够维持个体的多样性,后期能够加快算法的收敛速度,提高了DE算法的性能。最后对典型的稳定线性系统逼近问题进行了数值计算,计算结果证明该算法优于未改进微分进化算法,能够以更少的进化代数和更小的计算量找到高质量的逼近模型。

A Modified Differential Evolution （MDE） algorithm is proposed for realizing linear system approximation. The modified algorithm introduces an inertia scaling factor, which can dynamically maintain the diversity of the individuals at early stages and quicken convergence speed of the algorithm at later stages. Thus the performance of DE algorithm is improved. The approximation of a typical stable linear system was computed numerically, and the results showed that the algorithm is superior to the traditional DE, which can find a high-quality approximation model with less evolution generations and at lower computation cost.