Fuzzy Optimized Metric for Adaptive Network Routing

In this paper, a fuzzy inference rule base is implemented to generate the fuzzy cost of eachcandidate path to be used in routing the incoming calls. This fuzzy cost is based on the crispvalues of the different metrics; a fuzzy membership function is defined. The parameters of thesemembership functions reflect dynamically the requirement of the incoming traffic service as wellas the current state of the links in the path. And this paper investigates how three metrics, themean link bandwidth, queue utilization and the mean link delay, can be related using a simplefuzzy logic algorithm to produce a optimized cost of the link for a certain interval that is more‘precise’ than either of the single metric, to solve routing problem .