Efficient Resource Allocation for Multicell Heterogeneous Cognitive Networks With Varying Spectrum Availability

摘要

This paper investigates the problem of designing an efficient spectrum-allocation mechanism in heterogeneous multicell coordinated cognitive radio networks (CRNs) under time-varying channel availabilities and traffic demands across the different cells. The problem is modeled as a utility proportional fair maximization problem subject to traffic demand, spectrum heterogeneity, and exclusive spectrum sharing constraints. Specifically, the optimization problem is formulated as a binary linear programming (BLP) problem. Due to its integrality properties, the BLP is anticipated to be NP-hard. However, this paper proves that the optimal solution of the formulated BLP can be found in polynomial time using linear programming (LP) because of its total unimodular constraint matrix. The proposed solution is a spectrum-aware allocation design that ensures a fair spectrum sharing among cells based on their current traffic loads while considering the unique characteristics of their radio-frequency (RF) environment. Simulation results reveal that the proposed spectrum-aware proportional weighted fair allocation leads to improved system performance while effectively dealing with spectrum heterogeneity and time-varying traffic demands. The results also indicate that the proposed spectrum-allocation design is quite robust to traffic estimation and spectrum sensing errors.