Distributed joint power control, beamforming and spectrum leasing for cognitive two-way relay networks

Cognitive Radio (CR), as a promising technological solution to the spectrum under utilization problem, is becoming increasingly important as demand for various wireless applications and services rises. Protection of primary users from inflicted interference induced by the secondary users and, in the meantime, improvement of the network utility of the secondary users, thus remains the difficult challenge in underlay CR network. In the first part of the thesis, distributed power control and beamforming algorithm is proposed in which users operating in the underlay mode can strategically adapt their power levels and maximize their own utilities. This is subject to the primary user (PU) interference constraint as well as its own resource and target signal-tointerference- and-noise-ratio (SINR) constraints. The strategic competition among multiple decision makers is modeled as a noncooperative game where each secondary user (SU) acts selfishly to maximize its own utility. An adaptive method is proposed to determine appropriate pricing function. The problem of beamforming optimization under amplify-and-forward (AF) protocol is addressed as a generalized eigenvalue problem with respect to the utility function of SUs. The existence of a unique Nash equilibrium (NE) was proved and several numerical simulations were conducted to quantify the effect of various system parameters on the performance of the proposed method. In the second part of the thesis, maximization of the total revenue is formulated as an optimization problem that finds the optimal price and congestion threshold in the congestion-based pricing scheme. A search method with numerous advantages over conventional algorithms, has been designed to solve the optimization problems with an enhanced global optimality and convergence speed. Once the number of iterations conditioning along each dimension corresponds with the length of price interval, the convergence of algorithm is achieved. A key factor in the accuracy of Dynamic Response Pricing (DRP), the length of the demand response window, as observed in numerical results, indicates that the convergence of DRP to optimal threshold pricing is completed with a 98 percent accuracy in a 15 time-unit demand response. Finally, ADRP was introduced as an adaptive model of DRP, with numerical simulations of ADRP available for realistic call records data set. Simulation results show that optimal current channel occupancy and price is well tracked by ADRP. In order to know how much revenue may be lost because of the time-varying demand patterns, the first scenario was examined and the results show that 2734 monetary units per day for weekday were gained by the optimal threshold pricing, while this number for ADRP is 2652. In the second scenario, these values are 6595 and 6322 for optimal threshold pricing and ADRP, respectively. Comparing these results with optimal threshold pricing shows that ADRP loses just 3 percent and 4 percent of total revenue in first and second scenario, respectively.

Preview Text FK 2017 132 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item