Optimal Tradeoff Between RT and NRT Services in 3G-CDMA Networks Using Dynamic Fuzzy Q-Learning

With the evolution of wireless services and the increasing popularity of multimedia applications, setting radio parameters and sharing resources between different services are more than ever key issues for providing high-quality services. In this paper, we propose a dynamic and autonomic approach to sharing radio resources between RT and NRT services in order to achieve optimal trade-offs between their qualities of service. Fuzzy Q-learning controller (FQLC) is used to adapt the resource reserved for the RT traffic according to perceived quality indicators of each service. Each base station is fitted out with a controller which manages its resources. Learning results from each base station are supplied to a central FQLC of the network. Simulations have shown the effectiveness of the FQLC in terms of balancing the quality between RT and NRT services