QoS filtering in integrated wired/wireless networks using genetic algorithms

This paper presents a QoS filtering algorithm based on genetic algorithms (GA). The algorithm is used in integrated wired/wireless networks in order to improve the delivery of multimedia services. The proposed filters are located in every node of the network and are used to control adaptively the media content (i.e., video, voice, data, and image streams) downloaded from a media server to mobile users. The filtering algorithm performs a multi-objective optimization function that tries to meet the QoS requirements of different types of users (e.g. premium, or economy) by tuning the amount (as well as type) of the content to be delivered such that congestion is avoided. This action is performed by scaling the coding rate of each media type and performing, if necessary, a transcoding of some media, actually, downloaded. The filtering strategy provides a framework for an enhanced intelligent services platform by continuously adapting the contents of the services to the network time-varying conditions. This assures that critical QoS measuring parameters such as end-to-end maximum delay tolerance is always respected. Simulation results show that the optimization algorithm has significantly increased the number of admitted users while maintaining the users´ QoS requirements within acceptable levels.