基于差分演化二进制人工蜂群算法的多用户检测  

Multiuser detection based on differential evolution binary artificial bee colony algorithm

作　　者：刘婷[1,2] 张立毅[1,2] 邹康[3] 鲍韦韦[3] 

机构地区：[1]天津大学电子信息工程学院,天津300072 [2]天津商业大学信息工程学院,天津300134 [3]天津工业大学电子与信息工程学院,天津300387

出　　处：《电路与系统学报》2013年第1期5-10,共6页Journal of Circuits and Systems

摘　　要：最优多用户检测属于NP组合优化问题,人工蜂群算法作为一种简单有效的新兴启发式算法可以有效求解此类问题。针对基本二进制人工蜂群算法收敛速度慢、易陷入局部最优等缺陷,提出了一种基于差分演化的二进制人工蜂群算法,并应用于最优多用户检测中。算法采用多维邻域搜索策略,避免了连续域到离散域的转换,降低了算法复杂度,适合于实时处理。仿真结果表明,所提算法在抗多址干扰能力、抗"远近"效应能力和收敛性能方面均优于基本二进制人工蜂群算法。Optimum multiuser detection is a NP combinatorial optimization problem. As a simple, effective and new heuristic algorithm, artificial bee colony algorithm can effectively solve this problem. But the basic binary artificial bee colony algorithm has some defects such as slow convergence speed and being easy to fall into local optimum. A binary artificial bee colony algorithm based on differential evolution is proposed aiming at the above shortages. Multidimensional neighborhood search strategy is applied, and the conversion from continuous domain to discrete domain is avoided. The algorithm complexity is reduced, which is suitable for real time processing. The simulation results show that the algorithm＇s performances are better than the basic binary artificial bee colony algorithm in anti-multiple access interference, anti-near-far effect and convergence property.

关 键 词：最优多用户检测 基本二进制人工蜂群算法 差分演化二进制人工蜂群算法 邻域搜索策略 

分 类 号：TN914.53[电子电信—通信与信息系统]