Dynamic Spectrum Access in Time-varying Environment: Distributed Learning Beyond Expectation Optimization

摘要

This article investigates the problem of dynamic spectrum access forcanonical wireless networks, in which the channel states are time-varying. Inthe most existing work, the commonly used optimization objective is to maximizethe expectation of a certain metric (e.g., throughput or achievable rate).However, it is realized that expectation alone is not enough since someapplications are sensitive to fluctuations. Effective capacity is a promisingmetric for time-varying service process since it characterizes the packet delayviolating probability (regarded as an important statistical QoS index), bytaking into account not only the expectation but also other high-orderstatistic. Therefore, we formulate the interactions among the users in thetime-varying environment as a non-cooperative game, in which the utilityfunction is defined as the achieved effective capacity. We prove that it is anordinal potential game which has at least one pure strategy Nash equilibrium.Based on an approximated utility function, we propose a multi-agent learningalgorithm which is proved to achieve stable solutions with dynamic andincomplete information constraints. The convergence of the proposed learningalgorithm is verified by simulation results. Also, it is shown that theproposed multi-agent learning algorithm achieves satisfactory performance.