基于遗传算法的可重配置系统动态规划和管理  

Genetic Algorithm-Based Dynamic Planning and Management of Reconfigurable System

作　　者：徐玉滨[1] 丁哲[1] 沙学军[1] 

机构地区：[1]哈尔滨工业大学电子与信息工程学院,黑龙江哈尔滨150080

出　　处：《华南理工大学学报（自然科学版）》2010年第9期13-19,共7页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家"973"计划项目(2007CB310606);"新一代宽带无线移动通信网"国家科技重大专项(2009ZX03004-001)

摘　　要：可重配置无线通信系统的动态规划和管理是个复杂的组合优化问题.为此,文中提出一种基于遗传算法的方法,其能够在合法解空间中快速收敛到最优解,满足动态规划和管理的实时性要求.文中还提出了基于不同RAT频谱利用率差异的优化业务分布规划方法,以进一步提升系统的整体资源利用率.仿真结果表明,基于优化业务分布规划的动态规划和管理方法能够获得38.1%的频谱效率增益,相比基于传统业务分布规划的动态规划和管理方法大幅度提高了系统的整体频谱资源利用率.The dynamic planning and management（DPM） of reconfigurable system is regarded as a complicated combinatorial optimization problem.In this paper,a novel DPM method based on genetic algorithm is presented,which enables the fast convergence in the valid solution space and meets the real-time requirements of DPM.Moreover,a novol traffic distribution planning method based on the spectral efficiencies offered by different RATs is pre-sented to improve the overall spectral efficiency of the reconfigurable system.Simulated results demonstrate that the DPM method based on the proposed traffic distribution planning is of a spectral efficiency gain of 38.1%,and that it outperforms the one based on the conventional traffic distribution planning in terms of overall spectral efficiency.

关 键 词：动态规划 动态管理 遗传算法 可重配置系统 

分 类 号：TN925.5[电子电信—通信与信息系统]