An optimized protocol for cost effective communication in a multi-agent environment

Abstract

Mobile agents can migrate and communicate through interaction messages in a multi-agent environment. A multi-agent environment should provide a mechanism where mobile agents can be tracked easily and communication among them can take place while consuming the least amount of resources. Among all the proposed schemes the home based and forward pointing based scheme seems to be the best in achieving the desired objectives. The analysis shows that besides offering some advantages the aforementioned approach suffers from using high amounts of memory, high transmission cost, and the single point of failure problems. In this research we propose managing forwarding pointers intelligently by removing them from hosts when they are not required, preventing unnecessary update messages when the agent's mobility goes high, calculating the shortest path instantly after the agent's migration, and acknowledging only high priority interaction messages. We have validated our proposed approach with the help of experiments on a simulation tool OCEMAgents. The results show that the proposed approach has achieved all the objectives that are claimed and it has reduced memory and network overhead by minimizing the usage of forwarding pointers and by reducing the transmission cost. Besides that the proposed approach can effectively handle the single point of failure problem. Based on the results, we can safely declare that the proposed approach can be cost-effective if it is implemented in a real multi-agent system.