根据最小发送功率准则,针对多用户OFDM系统,提出了基于遗传算法的自适应比特加载解决方案.采用的矩阵编码方法直观、简单,具有并行性运算特性.通过交叉和变异让搜索跳离局部极值的陷阱,搜索到全局最优解.仿真结果表明,在给定的误比特率下,遗传算法与等比特分配方案相比可节省功率5~6dB,与multiuser adaptive OFDM(MAO)方法相比仅相差1dB,同时遗传算法的并行性使得节省了运算时间,提高了运算效率.
According to the criterion of minimum transmit power and multiuser OFDM system, an adaptive bit allocation scheme was presented on the basis of genetic algorithm (GA). The Matrix coding was visual and brief and characterized by parallel computing. The suboptimum was avoided and the global optimum solution was obtained by the crossover and mutation operation. The simulation results showed that genetic algorithm gained 5-6dB in comparison with equal bit allocation (EBA), and there was 1dB difference compard with multiuser adaptive OFDM(MAO). The parallel computing of genetic algorithm save running time and improved the calculation efficiency.