Using learning automata for adaptively adjusting the downlink-to-uplink ratio in IEEE 802.16e wireless networks

IEEE 802.16e allows for flexibly defining the relation of the downlink and uplink sub-frames´ width from 3:1 to 1:1, respectively. However, the determination of the most suitable ratio is left open to the network designers and the research community. Existing scheduling and mapping schemes are inflexibly designed. In this paper, a novel adaptive mapping scheme is proposed aiming to dynamically adjust the downlink-to-uplink ratio, following adequately the modification of the load requests with respect to both downlink and uplink directions. A learning automaton is exploited in order to sense the performance of the downlink and uplink mapping processes and to determine the most appropriate length ratio of both sub-frames in order to maximize the network performance. The suggested ratio determination scheme is evaluated through realistic scenarios and it is compared with static schemes that maintain a fixed ratio. The results show that our proposed scheme introduces considerable improvement, increasing the network´s service ratio and reducing the bandwidth waste.