An Adaptive Topology Management Algorithm in P2P Networks Based on Learning Automata

The topological structure of peer-to-peer networks is one of the topics of interest in these types of networks. Using the concept of community, as a technique for putting together the peers with similar interests, has largely contributed to the topological structure of peer-to-peer networks. A community is created when one or more numbers of a peer claim a similar interest about a common subject. Discovering a community and proposing it to a peer regardless of the priority of the keywords in the vector of interests leads to surplus connections and higher traffic in the network. Hence, in this paper, it is tried to, unlike previous studies and using learning automata, prioritize the interests of a peer in the vector of interest and the peer chooses more accurate communities for the sake of its own durability. Using this method, one node could choose those communities suggested to it with members with more similar interest. By implementing the interest-based searching method on the network obtained through the proposed algorithm, the search overhead and the success rate are investigated and the obtained results proved our claim.