Resource Allocation And Routing Selection Optimization In Cognitive Radio Networks

With the breakthrough of wireless network technology and the deep integration of Internet technology,the services carried by wireless networks have changed from simple voice calls to complex mixed services(for example,audio,video and data).On the one hand,it is not easy to guarantee users' Quality of Service(QoS)due to delays and bit error rates.On the other hand,resource utilization is low due to backward spectrum utilization strategies,severe user interference,channel fading,and noise.Therefore,how to improve the resource utilization rate while ensuring the user's QoS is an urgent problem.The secondary users(SU)of cognitive radio network(CRN)access the licensed spectrum of authorized users(primary user,PU)through the opportunity to achieve SU and PU spectrum sharing.CRN has broken the restriction of static spectrum access and improved spectrum utilization efficiency,which has attracted wide attention from the industry.At present,the research around CRN mainly focuses on the improvement of network performance and throughput.However,the flexibility of radio spectrum access brings new challenges to CRN,and at the same time adds complexity to the design of communication protocols at different layers.Spectrum allocation and data transmission are the key technologies of CRN,and these key technologies require energy consumption.Effective resource allocation schemes,routing strategies,and energy continuity guarantees are essential for emerging communication methods like CRN.This paper studies the two network scenarios of energy limitation and energy replenishment,and proposes corresponding resource allocation and routing algorithms to meet the QoS and energy consumption requirements of network users.First,for the energy-constrained CRN resource allocation optimization problem,this paper proposes a joint optimization algorithm for channel allocation and power control of CRN under multiple constraints.The problem of maximizing the total rate of SU nodes considering QoS,interference temperature and outage probability constraints is proposed,but this problem is a mixed integer nonlinear programming(MINLP)problem that is not easy to solve.To this end,the objective function is divided into two sub-optimization problems:channel allocation and power control.A genetic algorithm-based channel assignment algorithm(GACA)is proposed to obtain the optimal channel assignment strategy.The power control optimization problem is then a fractional form function with coupling constraints.For this reason,when SINR is high enough,the optimal power control strategy is obtained by introducing geometric programming(GP)and auxiliary variables to convert non-convex to convex geometric programming(CGP).When the SINR is at a low value,the iterative algorithm of single condensation method(SCM)is used to solve it.According to the characteristics of channel state information(CSI),a perfect CSI optimal solution and a non-perfect CSI suboptimal solution are designed.Simulation results show that the algorithm can achieve better performance in different CSI states.Secondly,in order to solve the shortcomings of CRN's energy limitation,this paper studies the power control optimization problem of CRN in the environment with energy harvesting(EH).The optimization goal is to maximize system capacity by allocating optimal power,while considering interference,SINR,energy saving,and QoS guarantees.In order to solve the non-convex nonlinear programming optimization problem,a Q-learning resource allocation algorithm(QLRA-EHCRN)based on reinforcement learning(RL)is proposed for the resource allocation of energy harvesting CRN.Through theoretical analysis and simulation experiments,it is proved that the algorithm can effectively increase system capacity,improve resource utilization,reduce power interference from SU nodes to PU nodes,and maximize system transmission rate.Thirdly,the routing problem of CRN in communication scenarios with energy supply is studied.The routing optimization process is modeled as a partially observable Markov decision process(POMDP).In order to find a routing strategy that meets the requirements of maximizing transmission rate and minimizing energy consumption,this paper proposes an energy harvesting routing algorithm(EHR-QL)based on Q-learning,which is used for multi-hop CRN routing selection.Numerical simulation results show that the algorithm can extend the network life cycle,increase the average throughput,and reduce the average end-to-end delay.Finally,this paper studies the problem of resource allocation and routing optimization of the CRN in energy harvesting communication scenarios.In order to improve the reliability of data transmission,we give the steps of the relay node selection,and also propose a hybrid game theory routing and power control algorithm(HGRPC)to solve the routing optimization problem with maximum system throughput.SU nodes on the same path cooperate with each other,and SU nodes on different paths compete with each other.By selecting the best next-hop node,you can find the best strategy to maximize throughput.Experimental results show that HGRPC has higher throughput,longer network life,less delay and lower energy consumption.