Achieving task allocation in swarm intelligence with bi-objective embodied evolution

Abstract In this paper, we seek to achieve task allocation in swarm intelligence using an embodied evolutionary framework, which aims generate divergent and specialized behaviors among a of agents online self-organized manner. our considered scenario, specialization is encouraged through bi-objective composite fitness function for the genomes, weighted sum local global function. The former depends only on behavior agent itself, while latter effectiveness cooperation all nearby agents. We have tested two existing variants evolution scenario compared their performances against those individual random walk baseline algorithm. found out that algorithms good at extreme cases weight configurations, but are not adequate when objective functions interact. thus propose novel algorithm, handles aforementioned by controlling proportion via dynamic reproductive isolation mechanism. Its other algorithms, as well theoretical Pareto frontier produced NSGA-II.