Evolutional Self-organization of Distributed Autonomous Systems

This paper discusses a framework and algorithms for reconfiguration of distributed autonomous systems (hereinafter called DAS). Most of DASs have ill-structured systems which mean nonlinear and complex systems. The proposed method based on Genetic Algorithms (hereinafter called GA) is applicable to optimize such a system. Many parameters of ill-structured systems are hierarchically mapped into parameters of the method as chromosomes in GA. The parameters of the system regarded as chromosomes are updated to increase (or decrease) fitness of each state of the system. As one example, the proposed method is applied to realize the self-organization of intelligent system of cellular robotic system (hereinafter called CEBOT) previously studied by authors. Both hardware and software of CEBOT consists of many functional units called “cells”. In this paper, self-organization of the intelligent system of CEBOT is mainly discussed. The proposed intelligent system consists of many kinds of knowledge sources, which have simple levels of data and intelligence together with some blackboards which dynamically organize a hierarchical structure. It is assumed that each intelligent unit is allowed to locally communicate with the other units to exchange data and information. Therefore, communication load of units can be defined. It is desired that the intelligent system can reorganize in order to maximizing (or minimizing) performance index calculated by using load of each unit. The efficiency of the proposed method is shown by some simulations of maximizing (or minimizing) the performance index of the system during self-organization.