Novel fuzzy non-dominance shortest path routing and path ordering for QoS aware routing

In this paper a novel fuzzy non-dominance routing and path ordering algorithm is presented for quality-of-service (QoS) routing for networks with inaccurate information. As one of the essential characteristic of both mobile wireless and fixed networks, the information available for decision making of QoS routing is always inaccurate[1][2]. Typically the link state information is collected by a decision maker as a result of dynamic traffic demand, channel quality, mobility, routing aggregation, out of date routing updates as well as any combination of the above factors. Therefore it is very difficult to model an exact cost for individual QoS path across the network. In this paper, we consider that network link weight is naturally uncertain due to inaccurate information. We define the routing problem in a fuzzy weighted graph. A novel concept of fuzzy shortest path and non-dominated multi-path will be introduced into network routing. A combination of fuzzy controller, linkstate fuzzy shortest path routing and path ordering algorithm is developed for finding QoS paths through the network with inaccurate information. It supports a better network QoS with less computational and management cost. We examine our fuzzy non-dominance routing scheme under a rate-based service model in[3]. Under such circumstance the available bandwidth is considered as a main parameter to guarantee the network QoS. We develop the simulation on OPNET Modeler and demonstrate our results based on randomized partial meshed mobile access networks. However this is easy to extend to other type of networks with different key metrics.