Adding Fault Tolerance to MaSE Methodology for Multi-Agent Systems

Many software applications today are designed based on multi-agent concept. These applications which need to work non-stop and continuos for a long period are prone to errors. Therefore, they are required to be fault tolerant as much as possible. Most of fault-tolerant methods provided for multi-agent systems are based on replicating the number of agents which leads to more complexity in these systems. In this paper we used task replication along with increasing numbers of agents in MaSE methodology to have a reliable architecture with high fault tolerante capability for multi-agent systems.