A conceptual framework for agent spawning

Abstract

Multi-agent systems are subject to performance bottlenecks in cases where agents cannot perform tasks by themselves due to agent overload, insufficient resources or network constraints. Solutions to such problems include distributing tasks to other agents or migrating to remote hosts. In this paper, we propose a conceptual framework for agent spawning, an alternative approach to the problem of local agent overloads. Agent spawning subsumes task transfer and agent mobility. Our paradigm entails that agents may spawn, pass tasks to others, merge or die. Spawning results in better software performance in terms of processing time, enhances efficiency of network utilization and reduce communication and computation loads. © 2014 IEEE.