Distributed evolutionary optimization, in Manifold: Rosenbrock's function case study

Abstract

A competitive coevolutionary approach using loosely coupled genetic algorithms is proposed for a distributed optimization of Rosenbrock's function. The computational scheme is a coevolutionary system of agents with only local interaction among them, without any central synchronization. We use a recently developed coordination language, called Manifold, to implement our distributed optimization algorithm. We show that this implementation outperforms a sequential optimization algorithm based on standard genetic algorithms.