ABSTRACT

Multi-Agent systems are virtual representations of social systems consisting of coarse grained intelligent entities, called agents. The motivation behind considering such paradigms is to emulate social behaviour among agents. Since an agent may not have the capabilities to accomplish all its tasks, it seeks cooperation of other agents in its society. Such goal directed cooperative efforts require communication among agents. In this paper, we present a scheme to manage communication in a large society of agents where there could be heavy communication traffic load due to all possible interactions among agents. The purpose of our communication scheme is two fold; (1) to manage communication traffic load by employing an agent partitioning strategy, (2) to maintain a behavioral discipline among agents by enforcing certain social laws during communication. Enforcement of social laws also ensures cooperative attitude among agents. The applicability of the proposed communication scheme is shown by taking an example from corporate sector.