Covariance matrix compact differential evolution for embedded intelligence

Abstract

This paper presents a compact evolutionary algorithm called the covariance matrix compact differential evolution (CMcDE). CMcDE is a real-parameter optimization evolutionary algorithm that adopt crossover in eigenvector space and representing population of search solutions as Gaussian probability distribution. The proposed algorithm has been tested using a standard test set of numerical problems. The experimental results show that the proposed CMcDE algorithm outperforms other algorithms in the test sets.