Restless multi-armed bandits under time-varying activation constraints for dynamic spectrum access
Kobi Cohen, Qing Qi Zhao, Anna Scaglione · 2014 48th Asilomar Conference on Signals, Systems and Computers · 2014
We consider the dynamic spectrum access problem where a cognitive user searches for transmission opportunities over a wide spectrum. We formulate the problem as a restless multi-armed bandit (RMAB) problem with a time varying constraint on the set of arms that can be activated at each given time. We consider the case where the channel occupancy dynamics are known, and we focus on establishing sufficient conditions on the optimality of the myopic policy.