An Improved Grey Prediction Evolution Algorithm Based on Topological Opposition-Based Learning

Abstract

The grey prediction evolution algorithm based on the even grey model (GPEAe) proposed by Z.B.Hu et al. in 2019 is a competitively stochastic real-parameter optimization algorithm with characters of simple code, less parameters and strong exploration capability. To improve the algorithmic overall performance, a topological opposition-based learning strategy (TOBL) is first developed to enhance its exploitation capability in this paper. The TOBL determines offsprings by calculating the Manhattan distances between the current best individual and all the vertices of the hypercube inspired by the opposition-based learning strategy. An improved grey prediction evolutionary algorithm based on the TOBL (TOGPEAe) is then proposed. The performance of the TOGPEAe is tested on CEC2005, CEC2014 benchmark functions and a test suite composed of six engineering design problems. The experimental results of the TOGPEAe are very competitive compared with those of the original GPEAe and other state-of-the-art algorithms.

Keywords

• Engineering design problems
• grey prediction evolution algorithm
• topological opposition-based learning