An Improved Annealing Algorithm Based on Multi-Agent

Abstract

Considering \u00a0the \u00a0characteristics \u00a0of Simulated \u00a0Annealing \u00a0Algorithm \u00a0(SAA), \u00a0based \u00a0on \u00a0the \u00a0agents\u2019 perception and retroaction to their surroundings, in this paper a new algorithm called Multi-agent Annealing Algorithm \u00a0(MAA) \u00a0is \u00a0presented. \u00a0And \u00a0the \u00a0global \u00a0convergence \u00a0of \u00a0MAA \u00a0is \u00a0gotten. \u00a0Finally, \u00a0we \u00a0make \u00a0several numerical \u00a0experiments \u00a0to \u00a0compare \u00a0MAA \u00a0with \u00a0SAA \u00a0and \u00a0Genetic \u00a0Algorithm \u00a0(GA) \u00a0by \u00a0three \u00a0classical \u00a0test functions in optimization, which suggest that MAA is superior to GA and SAA.