Heuristic Algorithms for Dynamic Spectrum Assignment in Open Spectrum System

DOI：

 作者 单位 郝丹丹 北京邮电大学 网络与交换技术国家重点实验室,北京 100876 邹仕洪 北京邮电大学 网络与交换技术国家重点实验室,北京 100876 程时端 北京邮电大学 网络与交换技术国家重点实验室,北京 100876

在回顾了现有的开放式频谱系统中的动态频谱分配算法后,基于快速收敛和公平性两方面的性能因素并兼顾系统总带宽性能,提出了两种易于实现且具有良好收敛性能的启发式频谱动态分配算法——兼顾最大化系统总带宽的快速收敛算法(fast convergency algorithm with maximum bandwidth,简称FCMB)和兼顾最大化系统总带宽的启发式公平性分配算法(heuristic fairness algorithm with maximum bandwidth,简称HFWB).通过大量的仿真实验,就系统总带宽、公平性以及收敛性能3个方面,与现有的协调式最大化系统总带宽(collaboration max-sum-bandwidth,简称CMSB)算法、随机分布式算法(randomized distributed algorithm,简称RAND)以及以最大化系统总带宽为目标的理论最优(theoretical max-bandwidth optimal,简称OPTL)算法进行了比较,并针对主、次用户数目变化、系统中信道数目以及次用户干扰区域半径大小变化等不同系统参数情况下各种算法的性能进行了对比分析.仿真结果表明,在综合考虑系统总带宽的基础上,FCMB算法和HFWB算法在快速收敛和兼顾系统带宽的公平性能上分别表现突出,尤其是FCMB算法,其在收敛速度上远远优于其他算法(和与其在系统吞吐性能上表现相近的CMSB算法相比,在收敛性能上至少有300%的提高).

Considering the factor of system sum bandwidth, two heuristic dynamic spectrum assignment algorithms for open spectrum systems are proposed according to convergency and fairness based on existing ones, they are the fast convergency algorithm with maximum bandwidth (FCMB) and the heuristic fairness algorithm with maximum bandwidth (HFWB). The performance of FCMB and HFWB is compared with the one of the collaboration max-sum-bandwidth (CMSB) algorithm, the randomized distributed (RAND) algorithm and the theoretical max-bandwidth optimal (OPTL) algorithm in system sum bandwidth, fairness and convergency by simulations. Furthermore, the effect of the numbers of primary users, secondary users and channels as well as the radius variance of disturbance area on the performance of those two algorithms is studied. Taking the system sum bandwidth into consideration, experimental results show that FCMB and HFWB outperform other three algorithms in convergency and fairness respectively, particularly FCMB shows superior performance in convergency (more than 300% improved than CMSB while FCMB performs similarly in system throughput).
HTML  下载PDF全文  查看/发表评论  下载PDF阅读器