摘要
提出了一种新的免疫算法,将匹配滤波器的输出判决信号作为一条染色体放入初始种群,使得初始种群中包含最优解的概率大大增加;选择疫苗时,将接种了疫苗的个体的适应值小于父代适应值的个体,按一定的概率接受该个体,克服了GA欺骗问题.仿真结果表明该算法在收敛速度和误码率方面都优于基于遗传算法的多用户检测.
A new method based on immune algorithm is proposed. By putting the decision of the output of a bank of matched filters into initial population as a chromosome, the probability that the best solution in included in the initial population increased. The individuals below fitness value of its parents are accepted in a probability, overcoming the trick problem in GA. Simulation results show that the proposed algorithm out-performs multiuser detection based on genetic algorithm in terms of convergence and BER performance.
出处
《甘肃科学学报》
2004年第3期58-60,共3页
Journal of Gansu Sciences
关键词
码分多址
多址干扰
多用户检测免疫算法
code division multiple access
multiple access interference
multiuser detection
immune algorithm
