Competitive multi-swarm system in adaptive resource allocation for a multi-process system

Abstract

This article presents a new proposal that performs adaptive resource allocation to control a multi-process system using bio-inspired techniques, based specifically on swarm intelligence algorithms (SI). These algorithms solve highly complex problems from simple rules to show adaptive and cooperative characteristics among various individual agents. This leads to observe with expectation the performance in the applicability that these algorithms handle to control complex systems with multiple inputs and multiple outputs. This paper presents a proposal of different swarm models run independently that recreate a competitive system from an operational standpoint, because each model is based only on its own assigned properties. This proposal was studied and analyzed using various performance metrics. All tested algorithms were able to control all assigned processes and especially the ant model showed more stability.