Self-organized criticality in differential evolution algorithm

Abstract

In this paper, we investigate self-organizing critical (SOC) properties of a successful evolution strategy Differential Evolution (DE) Algorithm. Numerical simulations of different DE strategies are studied on a set of problems. We search for typical signs, which point to a SOC behavior. Our main target is to find a power law-polynomial, scale invariant relationship. The power law is a basic characteristic of SOC systems, which can serve as a strong indicator of SOC.