Towards Adaptive WLAN Frequency Management Using Intelligent Agents

Private, corporate and public Wireless Local Area Networks (WLAN) Hot-Spots are emerging. In this rapidly evolving environment, the configuration of WLAN access points raise the classical problem of re-using limited radio resources. In this paper, the problem of dynamic frequency allocation of WLAN access point in a highly competitive multi-provider Hot-Spots environment is addressed. Our solution aims at working in locations where planned and ad-hoc deployments might be side by side. An on-line adaptive optimization process is proposed and relies on available information delivered only by the local access points in order to maintain the quality of service as high as possible. This optimization process is implemented on a scalable and highly flexible agent-based framework. The easy deployment of intelligent agents in a real WLAN network and their integration in a simulation context allows us to perform extensive tests for small and large-scale networks. The proposed approach has been tested on a limited but practical demonstrator that showed encouraging results.