基于粒子群优化的偏斜α稳定分布参数估计  

Parameter estimation for skewed α-stable distribution based on PSO

作　　者：李静威[1] 全厚德[1] 崔佩璋[1] 

机构地区：[1]军械工程学院光学与电子工程系,河北石家庄050003

出　　处：《电波科学学报》2013年第1期163-168,共6页Chinese Journal of Radio Science

基　　金：总装备部科研项目

摘　　要：基于粒子群优化算法,结合偏斜α稳定分布的分数低阶矩性质,提出了一种新的α稳定分布参数估计方法.针对已有方法无法实现多个参数同时估计的不足,结合α稳定分布分数低阶矩的特点,首先判断出对称参数的符号以及分数低阶矩阶数的取值范围,进而确定目标函数.仿真试验结果表明:所提方法可以同时给出3个参数的估值,验证了该方法的有效性,同时与Kuru-FLOM相比在精度上有所提高或相当.Based on particle swarm optimization（PSO）and the properties of frac- tional lower order moment（FLOM） of a stable distribution, a novel method of pa- rameter estimation for skewed a-stable distribution is proposed. Aiming at the dis- advantage of existing method without giving multi-parameter at the same time and utilizing characteristics of FLOM, the sign ofβ is detected and the moment of the FLOM is selected. Based on above/previous analysis the objective function is given. It is demonstrated by simulation results that three parameters can be obtained through PSO-FLOM directly. The novel method is effective and the accuracy is ad- vanced or equivalent to Kuru-FLOM.

关 键 词：分数低阶矩 粒子群优化 偏斜α稳定分布 参数估计 

分 类 号：TN911.7[电子电信—通信与信息系统]