Study On The Partical Swarm Optimization Algorithm And Its Application In PMD Compensation

Particle Swarm Optimization (PSO) is popular based stochastic intelligent optimization technique developed by Kennedy and Eberhart in 1995.Compare to genetic algorithm, the PSO algorithm hasn't depend upon the genetic operator to operate the individual, and has interacted information through own individual extreme value and the overall situation extreme value of the swarm. And it has much characteristic such as the simple of operation, easier realization and the quicker convergence rate and so on. It has already widely applied to function optimizes, nerve network and pattern recognition and so. Firstly, this article has introduced the background, research status and applications of standard PSO algorithm on detail, and pointed out the research direction and focus of the PSO algorithm, and expounded the basic principle, process and main features of the PSO algorithm.Secondly, due to the performance of PSO algorithm having largely depend upon the value of control parameters, this article mainly focuses on analyzing the relationship between convergence property of the PSO algorithm and the fixed parameter tuple {w , c₁ , c₂ }. Besides,it has made analysis about the parameter choice of the fixed parameter tuple.Thirdly, for the purpose of balancing the global exploration and local exploitation ability of the intelligent optimization algorithm, consider the fact that it is a non-linear move process for PSO algorithm to search the optima in the solution space, a new strategy is presented in this paper that the fixed parameter tuple {w , c₁ , c₂ } dynamically changes based on the run and evolution state. In addition, the simple PSO algorithm, which discards the velocity item, is more concise with guar detailedlyanteeing the convergence speed and accuracy. The modified adaptive simple particle swarm optimization (ASPSO) algorithm described in this paper is just based on the above two points.Finally, to compensate the effect of polarization mode dispersion (PMD) in fiber-optic transmission systems efficiently, the ASPSO algorithm was introduced into the PMD compensation system. This compensation system has many advantages, such as high sensitivity, short response time, little error and so on. Thus the compensation system has value of application in the improvement of transmission of optical fiber communication system.