A Reinforcement Learning Approach to Self-Configuring Edge Wireless Networks

Abstract

Wireless Internet access has brought legions of heterogeneous applications all sharing the same resources. However, current wireless edge networks that cater to worst or average case performance lack the agility to best serve these diverse sessions. Simultaneously, software reconfigurable infrastructure has become increasingly mainstream to the point that dynamic per packet and per flow decisions are possible at multiple layers of the communications stack. Exploiting such reconfigurability requires the design of a system that can enable a configuration, measure the network performance statistics (Quality of Service), learn the impact on the application performance (Quality of Experience), and adaptively select a new configuration. The goal of this work is to design, develop and demonstrate a reinforcement learning approach to self-configuring wireless edge networks that in instantiates this feedback loop. Our context is that of reconfigurable queueing, and we use the popular application of video streaming as our example. Through simulation and experimental validation, we show how measurement, learning and control are combined to enable high QoE video streaming on our platform.