This paper studies a generic model for cooperative cognitive radio networkswhere the secondary user is equipped with a finite length relaying queue aswell as a finite length battery queue. Our prime objective is to characterizethe stable throughput region. Nevertheless, the complete characterization ofstable throughput region is notoriously difficult, since the computation of thesteady state distribution of the two-dimensional Markov Chain (MC) model forboth finite queues is prohibitively complex. We first propose an algorithm tocharacterize the stable throughput region numerically, and show its sheercomputational complexity for large queue lengths. Hence, we next focus on twosimpler systems, namely, finite battery queue with infinite relay queue andfinite relay queue with infinite battery queue. The motivation behind therelaxation of having two finite queues at the same time is to lendtractability, explore the nature of design parameters optimization at thecognitive node and provide efficient lower computational complexity algorithmsfor stable throughput region characterization. For each proposed system, weinvestigate the maximum service rate of the cognitive node subject to stabilityconditions. Despite the complexity of the formulated optimization problems dueto their non-convexity, we exploit the problems' structure to transform theminto linear programs. Thus, we are able to solve them efficiently usingstandard linear programming solvers. Our numerical results demonstrate that, inpractical systems, finite battery and relaying queues achieve the same level ofbenefits of a system with infinite queue sizes when their sizes aresufficiently large. They also reveal that the achievable stable throughputregion significantly expands when the arrival rate of the energy harvestingprocess increases.
展开▼