Multichannel Opportunistic Spectrum Access in Fading Environment Using Optimal Stopping Rule

This paper studies the tradeoff between throughput and multichannel diversity in multichannel opportunistic spectrum access (OSA) systems. We explicitly consider channel condition as well as the activities of the primary users. We assume that the primary users use the licensed channel in a slotted fashion and the secondary users can only explore one licensed channel at a time. The secondary users then sequentially explore multiple channels to find the best channel for transmission. However, channel exploration is time-consumed, which decreases effective transmission time in a slot. For single secondary user OSA systems, we formulate the channel exploration problem as an optimal stopping problem with recall, and propose a myopic but optimal approach. For multiple secondary user OSA systems, we propose an adaptive stochastic recall algorithm (ASRA) to capture the collision among multiple secondary users. It is shown that the proposed solutions in this paper achieve increased throughput both the scenario of both single secondary user as well as multiple secondary suers.