基于NSGA-Ⅱ算法的认知无线电频谱分配  

Spectrum Allocation for Cognitive Radio Based on NSGA-Ⅱ

作　　者：王茵梦[1] 王松青[2] 杜庆伟[1] 

机构地区：[1]南京航空航天大学计算机科学与技术学院,江苏南京210016 [2]南京军区南京总医院,江苏南京210002

出　　处：《计算机与现代化》2015年第10期77-82,共6页Computer and Modernization

基　　金：国家自然科学基金资助项目(61202350)

摘　　要：认知无线电中,频谱拍卖是解决动态频谱分配的有效方法,其主要目的是最大化所有主用户的收益。然而,主用户间的收益是存在冲突的,给拍卖算法的优化带来了困难。为此,提出一种新的拍卖方案,来解决获胜者确定问题(Winner Determination Problem,WDP)。在该方案中,动态频谱拍卖问题被建模成多背包问题,并通过非支配排序遗传算法II(NSGA-II)得到最终的解决方案。最后,仿真实验结果表明,就解决WDP问题而言,与贪心算法相比,NSGA-II算法有更好的表现。Spectrum auction is an effective way to solve the dynamic spectrum allocation in cognitive radio networks. The main objective in an auction is to maximize all the primary users＇ revenues. However,there are some conflicts among primary users＇ gains and this brings difficulties to do the auction. So,a new auction scheme is proposed to solve the winner determination problem（ WDP）. In the scheme,the dynamic spectrum allocation is regarded as multi-knapsack problem and trying to solve it by non-domination sort genetic algorithm II（ NSGA-II）. Finally simulation experiments show that NSGA-II does a good job in solving the problem of WDP compared with the greedy algorithm.

关 键 词：认知无线电 动态频谱分配 频谱拍卖 NSGA-Ⅱ 多背包问题 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]