Research On Optimization Models And Algorithms In Cooperative Cognitive Radio Networks

Nowadays, with the ever-increasing requirement of the network bandwidth and the emergence of diverse wireless network applications, it has become challenging to still use the traditional fixed spectrum allocation strategy. To deal with this problem, cognitive radio network is proposed, which is a joint technology that incorporates cognitive technology, computer technology, information technology, and control technology, etc. Since it can realize efficient spectrum utilization through sensing and decision-making process to the wireless environment, and has become a hot topic both in academia and industry. In order to further exploit the performance of cognitive radio networks, cooperative technologies such as cooperative sensing and cooperative transmission can be adopted. Thus a new paradigm- cooperative cognitive radio networks has emerged. In this thesis, we are motivated to investigate several open problems of cooperative cognitive radio networks. The main contributions are summarized as follows.(1) Above all, we give a survey to cooperative cognitive radios networks, which includes its categories, fundamental technologies, and current hot research topics. We then point out the weakness of the current works and propose some open problems.(2) To meet the requirement for the energy-efficient cooperative sensing application in cognitive radio networks, we address two sub problems: the energy minimization node selection problem (EMNS) and the online energy-efficient node selection problem (OENS). We first prove their NP-hardness by reducing them to the well-known two-dimensional cost knapsack problem and the disjoint connected path problem, respectively. Then, for EMNS, we propose a branch and bound algorithm (BAB) to find its optimal solution, and design a heuristic greedy selection algorithm (GS) to get an approximate solution. For OENS, for each node, we introduce a weight coefficient which takes into account the load balancing of its energy consumption, then based on algorithm BAB and GS, we propose two online node selection algorithms OBAB and OGS1. Extensive experiments by simulations demonstrate that OGS1 can solve the OENS problem effectively. In terms of the total accomplished rounds of successful cooperative sensing progress, OGS1 can deliver a solution that is up to 94% of the near optimal solution gained by OBAB, meanwhile decreasing the algorithm running time significantly.(3) Cooperative sensing technology can greatly improve the spectrum utilization in cognitive radio networks. However, with the formation of cooperative sensing coalition, it inevitably introduces extra cost. All the nodes within a coalition expect to achieve a higher throughput with less extra energy cost. We thus address a multi-objective non-linear programming problem for cooperative sensing. Based on coalition game theory, we construct a non-transferable utility coalition formation game for the problem. In the design of its payoff function, we assign the throughput expectation and the energy cost with different weights, so that these two objectives are jointly considered. After that, we propose a distributed multi-objective coalition formation algorithm (DMCF), in which coalitions are iterated merged and split according to Pareto order. In addition, we show the convergence of the proposed algorithm and the stability of the final coalition partition. Extensive experiments by simulations demonstrate that, compared with the result delivered by a distributed random coalition formation algorithm (DRCF), algorithm DMCF can increase the node's expected throughput by 7.5%, meanwhile decreasing the energy cost significantly by 70%, which shows great effectiveness to deal with the proposed multi-objective optimization problem.(4) We propose a relatively robust optimization model under the scenario where multi-channels are cooperatively sensed and used by multi-SUs. The model aims to maximize the system throughput and optimizes the parameters including the sensing time and the weight coefficient of the sampling result of each SU for each channel, meanwhile the false access probability for each channel must not violate the given constraints. To solve this non-linear optimization model, we propose a sequential parameter optimization method (SPO). The method first optimizes the weight coefficients by solving a series of sub-optimal problems through Lagrange method. Then it transforms the problem into another monotonic programming problem and exploits a fast polyblock algorithm (FP) to search an optimized sensing time parameter. Extensive experiments by simulations demonstrate that, in terms of the throughput gained by the system, SPO can deliver a solution that is up to 99.3% of the optimal on average, which can effectively solve the proposed optimization model. In addition, the advantage of the proposed optimization model is further verified by comparing with other models.(5) Consider a real yet complex scenario where multi-primary users and multi-secondary users coexist and multi-secondary users and cooperate with a certain primary user, we are motivate us to propose a joint spectrum allocation and cooperation set partition problem, which so far has not been addressed before. We formulate the problem as a 0-1 integer non-linear programming model. Due to its NP-hardness, we propose a suboptimal centralized genetic algorithm (CGA) and show its convergence by modeling it as a homogeneous finite Markov chain. We then extend CGA to a fully distributed genetic algorithm (DGA) that consists of two phases. The core techniques include a minimum dominate set based cluster partition and spectrum pre-allocation algorithm in phase 1, and an inter-cluster cooperation set negotiation and cluster fitness refinement algorithm in phase 2. We also devise a Fast-convergent DGA (FDGA) to reduce the system configuration time. Extensive experiments by simulations demonstrate that in terms of the fitness that reflects the performance of the proposed algorithms, CGA is shown to perform as well as 92% of the optimal solution by brutal search under small network sizes. As the network size increases, due to the massive search space CGA has to deal, DGA and FDGA instead outperform CGA with 20% on average when achieving the same algorithm termination condition.(3) FDGA delivers similar results as DGA while reducing the configuration time significantly, which is more suitable for large-scale networks.In summary, this thesis investigates several open problems and provides theoretical optimization model and solution methods to further exploit the performance of cooperative cognitive radio networks. Our works have their academic and practical value on promoting the advancement of the researches and applications in cooperative cognitive radio networks.