Study On The Control Algorithm For Polarization Mode Dispersion Compensation In Optical Fiber Communication Systems

In recent years, to meet the need of communication capacity, the bit rate of optical fiber communication is improving all the time. The polarization mode dispersion (PMD) which was not serious in low bit rate optical fiber communication becomes the main signal distortion factor for high bit rate optical fiber communication. Although the particle swarm optimization is a comparatively good algorithm for PMD compensation, the noise in the optical communication system makes the algorithm not satisfied, and needs to be modified.Particle Swarm Optimization algorithm is a direct search algorithm has been widely used in various types of optimization problem. The comprehensive learning particle swarm optimization (CLPSO) and the adaptive comprehensive learning particle swarm optimization with history learning (AH-CLPSO ) which we will study in this thesis are two sorts of improved PSO algorithms and use in adaptive PMD compensation system. They improve the iterative equation of the particles' velocity, as a result, they have improvement advantage of fast convergence to the global optimum rather than the sub-optima. The main task of this thesis is to get an improved PSO algorithm and to test its performance in the control unit of PMD compensator, finally elected performance better algorithm applied adaptive PMD compensation system control module. The work in this thesis is summarized as follows:The program code of the CLPSO algorithm is given and its performance is tested. The result reflects that this algorithm converges fast and will not be trapped into sub-optima. The only disadvantage that the algorithm has is the slower running speed than the PSO algorithm.The program code of the AH-CLPSO algorithm is given and its performance is tested.The result reflects that this algorithm converges slower than the CLPSO algorithm and also will not be trapped into sub-optima. Due to its more complicated construction, it runs slower than the CLPSO algorithm.The DOP map which we get from experiments is treated using the CLPSO and AH-CLPSO algorithm to test the efficiency of the improved PSO algorithms. The result reflects that the CLPSO algorithm is better than the AH-CLPSO algorithm. It converges faster and will not be trapped into sub-optima and is easier to find the optimum. Also it runs faster than the AH-CLPSO algorithm.The two Algorithms will be used in the (DOP) "map" which is researched from PMD experiment .the test results show that: CLPSO performance better, it is faster, and easier to find optimal value.