摘要
文章介绍了复合重要抽样的基本理论,针对一阶偏振模色散模拟器仿真的不同配置,讨论了重要抽样的不同的偏差分布。仿真结果证明,采用复合重要抽样技术,不但可以获得概率为10-17的极小概率事件,与传统的蒙特-卡洛仿真方法相比,还大大提高了仿真效率。同时,通过对比重要抽样中不同的偏差分布函数的仿真结果,给出了不同系统配置所选取的不同的偏差分布。
A Multiple Importance Sampling （MIS） theory for first order Polarization Mode Dispersion （PMD） is introduced. Different biasing distributions are discussed on the basis of different configurations in first order PMD simulations. The simulation results show that by the MIS technique, extremely low probability （10^-17） events can be obtained and the simulation efficiency is greatly improved as compared with the standard Monte-Carlo techniques. Furthermore, by contrasting the simulation results of different biasing distribution functions in importance sampling, different biasing distributions for the different system configurations are given.
出处
《光通信研究》
北大核心
2007年第2期18-21,共4页
Study on Optical Communications
基金
西安邮电学院青年教师基金资助项目(101-0415)
