求解多输入多输出检测新算法——遗传粒子群优化

New genetic particle swarm optimization evolutionary algorithm for MIMO detecting system

设计两种基于粒子群优化算法(PSO)和基于遗传算法(GA)的多输入多输出(MIMO)系统检测算法。提出一种新的融合GA和PSO进化机制的遗传粒子群进化(GPSO)算法,并将其应用于MIMO系统检测问题求解。新算法改善了初始化种群,并将每一代粒子划为精英粒子、次优粒子和糟糕粒子三部分,对这三种粒子分别采用极值扰动、PSO进化和淘汰策略以改善算法的全局和局部搜索能力,从而加快算法的寻优速率和收敛速度。仿真结果表明:与基于PSO和基于GA的检测算法相比,GPSO的检测算法能够很大程度减少种群规模和迭代次数。而与最优的最大似然译码算法相比,GPSO检测算法能够在计算复杂度和误码性能之间获得很好的折中。

Two MIMO detecting algorithms corresponding to particle swarm optimization （PSO）based and genetic algorithm （GA）based detecting algorithms are designed. A novel genetic particle swarm optimization （GPSO） evolutionary method is proposed and applied to address the MIMO detecting problem. The proposed algorithm starts from improving the initial population, and divide the entire population into three types; elite particles, better particles and worst particles. Three different strategies of optimum value permutation, PSO evolvement and elimination strategy are employed corresponding to these three type particles to improve the local and global search ability. Therefore, both the optimum searching ability and the convergence speed are accelerated. Simulation results reveal that GPSO-based detecting algorithm takes much less size and less iteration number when compared with the PSO-based and the GA-based detecting method. Besides, compared with optimal maximum likehood（ML） detecting method, the GPSO-based detecting al- gorithm can reach better balance between the BER performance and the computational complexity.