2018/11/27 by Semih Çaycı, Cayci, Semih, Atilla Eryılmaz +1
Computer Science · Decision Sciences · #Cognitive Radio Networks and Spectrum Sensing #Age of Information Optimization #Advanced Bandit Algorithms Research
paper · pdf · doi:10.48550/arxiv.1811.10829
We study the problem of serving randomly arriving and delay-sensitive traffic\nover a multi-channel communication system with time-varying channel states and\nunknown statistics. This problem deviates from the classical\nexploration-exploitation setting in that the design and analysis must\naccommodate the dynamics of packet availability and urgency as well as the cost\nof each channel use at the time of decision. To that end, we have developed and\ninvestigated an index-based policy UCB-Deadline, which performs dynamic channel\nallocation decisions that incorporate these traffic requirements and costs.\nUnder symmetric channel conditions, we have proved that the UCB-Deadline policy\ncan achieve bounded regret in the likely case where the cost of using a channel\nis not too high to prevent all transmissions, and logarithmic regret otherwise.\nIn this case, we show that UCB-Deadline is order-optimal. We also perform\nnumerical investigations to validate the theoretical findings, and also compare\nthe performance of the UCB-Deadline to another learning algorithm that we\npropose based on Thompson Sampling.\n