Research On Spectrum Allocation Technology Based On Evolutionary Algorithm In Cognitive Radio Networks

With the rapid development of wireless communication and the continuous innovation of communication products,the demand for the spectrum resources is growing rapidly.The traditional distribution method is no longer applicable.Dynamic spectrum allocation technology can improve spectrum utilization to a certain extent but can not fundamentally solve the problem of scarce spectrum resources in the face of endless growth in demand.Cognitive radio(CR)is considered as a software radio technology that can fundamentally solve the spectrum scarcity problem.Dynamic spectrum sharing technology,as one of the most important technologies,allows secondary users to reuse the licensed spectrum on the premise of not affecting the primary users.Therefore,CR dynamic spectrum sharing technology is undoubtedly an important research topic.In this paper,The CR spectrum and power allocation strategy in the spectrum sharing system is investigated to maximize the throughput and fairness of the system and minimize the total transmission power.Firstly,relevant theories on CR spectrum sharing and power allocation are elaborated.And then the spectrum sharing model is put forward in this paper:spectrum allocation under a given matrix,available benefit matrix and interference matrix condition aiming at maximizing the total system reward,the minimum secondary user reward,and the proportion fairness.Further,the second spectrum sharing model,i.e.,joint spectrum and power allocation under the condition of limited transmission power in CR network,is then proposed.The theoretical analysis shows that the optimization problems are non convex optimization problems,which can be solved by evolutionary algorithms.Therefore,the improved chemical reaction optimization algorithm(ICRO)has been adopted in this paper.ICRO is one adaptive algorithm which integrates the excellent properties of the chemical reaction optimization(CRO),the particle swarm algorithm(PSO)and the differential algorithm(DE).Simulation results show that ICRO can perform better than other evolutionary algorithms.Finally,since the spectrum allocation scheme with single target may not meet the actual requirements,multiple targets always need to be considered.The third proposed spectrum sharing model is a multi-objective optimization problem under the limits of signal-to-noise in a link and the transmission power of each secondary user based on Underlay mode.The theoretical analysis shows that the multi-objective optimization algorithm is more effective than the weighted single objective algorithm.Additionally,the ICRO is extended to multi-objective case and the corresponding multiple objective chemical reaction(MICRO)is put forward to solve this problem.Simulation results show that MICRO is able to obtain much better Pareto solutions than other algorithms.