Self-Adaptive WLAN Access Point for Optimizing Network Performance Using Multi-Objective Genetic Algorithm (MOGA)
Authors: Joel C. Delos Angeles and Elmer P. Dadios
Abstract
Current deployment of WLAN access points (AP) require manual configuration of wireless parameters. Wireless parameters are commonly set haphazardly without being aware of the basic wireless conditions. This paper proposes a selfadaptive AP based on genetic algorithms (GA). The AP adapts to interference and link quality of client stations. Interference is mitigated and client link quality is improved or optimized. A chromosome consists of genes of parameters such as frequency channel, channel width, maximum data rate, maximum transmit power, and guard interval. Often competing objectives such as mitigating interference, maximizing the data rate, and minimizing the error rates necessitate that the GA be multi-objective. The MOGA comes up with the fittest candidates by running them through a fitness function which scores the genes based on the survey scan of other interferer AP and the wireless performance statistics of client devices. The GA’s chosen configuration is applied and its effect is continuously assessed. Finally, the result of the self-adaptive WLAN AP genetic algorithm is compared against the Linux hostapd Automatic Channel Selection scheme.