Multi-Agent Computer Network Routing System Based on Recursive Modeling Method and Bayesian Belief Update

This research addresses computer network routing based on recursive modeling method (RMM). In our proposed method, each router as an intelligent agent uses RMM tree-like structure to make a rational and intelligent decision on transmitting packages. This decision is made based on the models of other routers in the network and its belief about them. However, as the network conditions vary with time, the needs of updating each routers belief about others can not be neglected. To achieve this goal, Bayesian law is used to update routers belief about others. To validate the proposed method, we test it as a computer simulation. The results show that this method increases the network performance by decreasing the network overhead arising from routers packages. Additionally, as the network dynamics increases, employing Bayesian law helps the routers to quickly adapt to the new conditions.

کلیدواژه ها:

Computer Network Routing ، Recursive Modeling Method ، Multi-Agent ، Bayesian Belief Update