采用多策略离散人工蜂群的改进频谱分配算法  

An Improved Spectrum Allocation Algorithm Using Multi-Strategy Discrete Artificial Bee Colony Technology

作　　者：朱冰莲[1] 朱方方[1] 段青言 张良[1] 肖欣庭 

机构地区：[1]重庆大学通信工程学院,重庆400044

出　　处：《西安交通大学学报》2016年第2期20-25,84,共7页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(61201177)

摘　　要：针对图论频谱分配模型下最优频谱分配策略搜索解困难、耗时长的问题,提出一种采用多策略离散人工蜂群的频谱分配算法。首先,根据感知技术得到的通信环境状况,建立频谱分配的图论模型;然后,引入多策略离散人工蜂群算法进行最优频谱分配策略的搜索,在搜索初期,引入全局探索能力强的粗搜索策略,以快速优化初始种群,后期以高精度的单维更新进行精细搜索;考虑到仅当解参数值取1才能带来网络收益的增加,提出仅对取值为零的维度进行更新的策略,增强了搜索的有向性与有效性。仿真实验表明：该算法与当前基于离散人工蜂群和二进制粒子算法的频谱分配算法相比,无论是收敛速度还是网络收益都得到提高;当可用频谱数在5-20之间、次用户数量在5-22之间变化时,获得相同最大收益的耗时仅为对比算法的47.75%-36.18%,且随着问题规模增加耗时呈下降趋势。An improved spectrum allocation（MDABC-SA）algorithm using the multi-strategy discrete artificial bee colony technology is proposed to reduce computational time of spectrum allocation based on graph model.First,a spectrum allocation model is established based on parameters obtained by sensing technology.Then,the multi-strategy discrete artificial bee colony technology is employed to find the optimal spectrum allocation scheme,and a global searching operator is used in initial searches to rapidly find a better initial population,An one-dimensional search is then used in later searches to perform fine line search.The strategy to update only the elements with value of 0is proposed to inhance the direction and effectiveness of searches by considering the fact that the more‘1＇have in the solution,the higher network utilization can be achieved.Simulation results and comparisons with the spectrum allocation algorithms using DABC and BPSO algorithms show that the proposed algorithm obviously improves both the convergence speed and network utilization.The algorithm achieves the same maximum benefit with only 47.75%-36.18% of consumed time of the former two algorithms when the number of available spectrum is between 5and 20 and the number of secondary users varies from 5to 22 and a downward trend in consumed time is observed when the problem scale increases.

关 键 词：频谱分配 图论模型 人工蜂群 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TN925[自动化与计算机技术—控制科学与工程]