A modified Alopex-based evolutionary algorithm and its application on parameter estimation

In order to improve the efficiency of an Alopex-based evolutionary algorithm (AEA), a modified AEA algorithm (CAEA) which combines copula estimation of distribution algorithm (copula EDA) is introduced in this paper. In view of the inefficiency and the lack of adequate evolutionary information for the population in AEA, a set of competitive and elite solutions are acquired to improve the quality and maintain the diversity of the candidate population by using EDA based on copula. The modified algorithm not only takes advantage of heuristic search of AEA, but also inherits the superiority of rapid convergence of copula EDA. Then 22 benchmark functions are employed to test the performance of CAEA algorithm. Compared with AEA, EDA and differential evolution (DE), the optimization results indicate that the performance of CAEA is significantly superior to that of the other three algorithms, no matter in accuracy or in stability. Furthermore, the modified algorithm is applied to estimating the parameters of a fermentation dynamics model. The results of the comparisons with the other two algorithms illustrate that CAEA algorithm is effective in practical engineering application.