Throughput Satisfaction-Based Scheduling for Cognitive Radio Networks

Abstract

Opportunistic scheduling algorithms in cognitive radio networks (CRNs) allocate resources by exploiting the variations in channel conditions and spectral activities of primary users. However, most of these scheduling algorithms ignore the per-user throughput requirements. In this paper, we formulate a scheduling problem called maximizing the number of satisfied users (MNSU), which maximizes the number of secondary users that are satisfied in terms of throughput in a centralized CRN. We show that MNSU is NP-hard in the strong sense and cannot be approximated within any constant factor better than 2 unless <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$P = NP$</tex> </formula> . We also prove that MNSU is at least as hard as the max–min fair scheduling problem, which has previously been proven to be a computationally very difficult problem in the literature. We then propose two heuristic algorithms: 1) best first resource assignment and 2) resource assignment with partial backtracking. We demonstrate that our proposed algorithms yield high performance while still achieving low computational complexity.