A simplified non-equidistant grey prediction evolution algorithm for global optimization

Abstract

Grey prediction evolution algorithm, which is receiving more and more attention, considers the population sequence of evolutionary algorithms as an equidistant time series. In fact, the average fitness value of each generation population decreases at a variable speed during almost any evolution process. Therefore, the population sequence with the property of variable speed evolution should be modeled more properly as a non-equidistant time series. Along this way, this paper proposes a novel evolutionary algorithm based on a non-equidistant grey model, called simplified non-equidistant grey prediction evolution algorithm. The proposed algorithm is identified by its reproduction operator which is developed by the following three steps. Firstly, a non-equidistant grey model based on the average fitness value of each generation population is modeled. Secondly, the interval in the fitting stage of the non-equidistant grey model is simplified to an approximately equidistant time interval. Finally, a simplified reproduction operator which employs a parameter to preserve the non-equidistant nature in the prediction stage of the non-equidistant grey model is developed. The comprehensive performance of the proposed algorithm is evaluated on CEC2014 and CEC2019, as well as six engineering constrained design problems. Experimental results show that the proposed algorithm ranks first in CEC2014 and CECE2019 benchmark functions. The proposed algorithm can achieve a better solution with fewer computational overhead than some state-of-the-art algorithms on six engineering constrained design problems.The Matlab code of this paper is available onhttps://github.com/Zhongbo-Hu/Prediction-Evolutionary-Algorithm-HOMEPAGE.