Fuzzy Based Node Disjoint QoS Routing in MANETs by Using Agents

Abstract — Support for real time multimedia applications such as, video telephony, financial stock quote services, and multiplayer interactive games etc., is very essential in Mobile Ad hoc Networks (MANETs). Such applications require multiple Quality of Service (QoS) parameters to be satisfied, like bandwidth, end-to- end delay, packet loss rate, jitter, etc. This paper considers the problem of finding node disjoint and multi-constrained QoS multipaths from source to destination by using agent based fuzzy inference system. The proposed scheme, Fuzzy based Node Disjoint Multipath QoS Routing (FNDMQR) operates in the following steps by integrating static and mobile agents. (1) Determination of multiple paths and picking up of resource information (available bandwidth, link delay, and packet loss rate) of the intermediate nodes from source to destination. (2) Recognition of node disjoint, and multi-constrained QoS fit paths by using Takagi-Sugeno Fuzzy Inference System (TSFIS). TSFIS extracts a fuzzy QoS weight from available resource information of the intermediate nodes. (3) Selection of the best path depending on the fuzzy QoS weight. (4) Maintenance of QoS path when path breaks due to mobility of node or link failure. To test the performance effectiveness of the approach, we have analyzed the performance parameters like packet delivery ratio, average end-to-end delay and overall control overhead. The scheme performs better as compared to a node-disjoint multipath routing in MANETs. Index Terms—MANETs, QoS, Takagi-Sugeno fuzzy Inference, software agents