在前馈实验中加入扰偏器,利用粒子群优化算法以有限的实验采样数据快速拟合出了精确的偏振度椭球,实验分析了不同邻居结构的粒子群算法拟合偏振度椭球的有效性,证明在拟合偏振度椭球方面,局部邻居结构的粒子群优化算法要比全局邻居结构的算法性能更好,拟合出的椭球更准确,更适合于前馈补偿方案。
By adding a polarimeter into the experiment,exact degree of polarization (DOP) ellipsoids are obtained rapidly from the sampling data by using particle swarm optimization (PSO) algorithm. The performances of PSO with different neighbour structures are analyzed in the course of getting DOP ellipsoids, Results show that the ellipsoids obtained by the local version of PSO (LPSO) are more exact than that by the global version of PSO (GPSO), and LPSO is excellent fit for the feedforward compensation in getting exact ellipsoids.