Complex Algorithm Network and Its Toplogy Analysis: a Case Study

Complex networks are everywhere. This paper explores a new type of complex network, complex algorithm network constructed from evolutionary algorithms. The main aim of this paper is to investigate the topological properties underling the dynamics of the evolutionary algorithms. It uses complex network to describe individuals and their relationships while execution of a specific evolutionary algorithm. It introduces the main parameters in complex network theory to uncover properties of the dynamics of evolutionary algorithms. Our simulations are based on one selected evolutionary algorithm (Guo' algorithm in 10 versions) and one test functions. Data obtained through the simulations are processed graphically as well as statistically. Some novel properties of the complex algorithm network are uncovered, which has great implications for the design and improvement of existing evolutionary algorithms.