Cognitive forwarding control in wireless ad-hoc networks with slow fading channels

摘要

We propose a decentralized stochastic control solution for the broadcast message dissemination problem in wireless ad-hoc networks with slow fading channels. We formulate the control problem as a dynamic robust game which is well-justified by two key observations: first, the shared nature of the wireless medium which inevitably cross-couples the nodes' forwarding decisions, thus binding them together as strategic players; second, the stochastic dynamics associated with the link qualities which renders the transmission costs noisy, thus motivating a robust formulation. Given the non-stationarity induced by the fading process, an online solution for the formulated game would then require an adaptive procedure capable of both convergence to and tracking strategic equilibria as the environment changes. To this end, we deploy the strategic and non-stationary learning algorithm of regret-tracking, the temporally-adaptive variant of the celebrated regret-matching algorithm, to guarantee the emergence and active tracking of the correlated equilibria in the dynamic robust forwarding game. We also make provision for exploiting the channel state information, when available, to enhance the convergence speed of the learning algorithm by conducting an accurate (transmission) cost estimation. This cost estimate can basically serve as a model which spares the algorithm from extra action exploration, thus rendering the learning process more sample-efficient. Simulation results reveal that our proposed solution excels in terms of both the number of transmissions and load distribution while also maintaining near perfect delivery ratio, especially in dense crowded environments.