A predictive measurement-based fuzzy logic connection admission control

This paper presents a novel measurement-based connection admission control (CAC) which uses fuzzy set and fuzzy logic theory. Unlike conventional CAC, the proposed CAC does not use complicated analytical models or a priori traffic descriptors. Instead, traffic parameters are predicted by an on-line fuzzy logic predictor (Qiu et al. 1999). QoS requirements are targeted indirectly by an adaptive weight factor. This weight factor is generated by a fuzzy logic inference system which is based on arrival traffic, queue occupancy and link load. Admission decisions are then based on real-time measurement of aggregate traffic statistics with the fuzzy logic adaptive weight factor as well as the predicted traffic parameters. Both homogeneous and heterogeneous traffic were used in the simulation. Fuzzy logic prediction improves the efficiency of both conventional and measurement-based CAC. In addition, the measurement-based approach incorporating fuzzy logic inference and using fuzzy logic prediction is shown to achieve higher network utilization while maintaining QoS