An Optimal Architecture of Middle Agents in Multi-Agent Systems

The research about multiple middle agents chiefly focus on a variety of different architectures to increase system flexibility nowadays, but the optimal design to the architecture is very rare. This paper proposed a optimal logical ring architecture of multiple middle agents in multi-agent systems, according to probability theory, the system are simplified by reducing the numbers of middle agents under the condition of the system is not broken down with definite probability. The optimization results show that the stability and flexibility of the system don´t decrease although reducing the number of middle agents, moreover, middle agents can accomplish more tasks in the logical ring. Analysis results show that the presented method is effective and feasible.