基于多智能体粒子群算法的盲均衡技术研究

Research of Blind Equalization Technology Based on Multiagent Partical Swarm Optimization

在解决盲均衡的问题中,当前主要采用梯度搜索的方法,这种方法的优点在于局部最优解收敛,降低通信性能。本文根据粒子群算法,设计了各种算子,提出了用粒子群算法的快速搜索全局最优解的性能来优化盲均衡算法的代价函数,从而提出了一种新的基于粒子群算法的盲均衡算法。仿真结果表明,所提出的算法在较少的迭代次数情况下,能够快速收敛,具有较优的性能。

At present, in resolving the problem of blind equalization, we mainly used the gradient search method. This method has the advantage of the convergence of local optimal solution, but it reduce the communication performance. Based on Particle Swarm Optimization （ PSO）, we designed a variety of operator and proposed using PSO quick search global optimal solution to optimize the cost function of blind equalization. Simulation results show that the proposed algorithm was able to fast convergence with less number of iterations than the gradient search method algorithm under PSO circumstances and the communication performe of the proposed algorithm is also more better in its practical application.