重新构造L-M迭代参数,即μk=θ‖F k‖+(1-θ)min{‖F k‖,‖JT k F k‖},θ∈n[0,1],来求解非线性方程组F(x)=0.在算法中,当试探步不成功时,采取新的非精确线搜索技术获得下一个迭代点.在适当假设条件下,证明了该算法具有全局收敛性....重新构造L-M迭代参数,即μk=θ‖F k‖+(1-θ)min{‖F k‖,‖JT k F k‖},θ∈n[0,1],来求解非线性方程组F(x)=0.在算法中,当试探步不成功时,采取新的非精确线搜索技术获得下一个迭代点.在适当假设条件下,证明了该算法具有全局收敛性.数值实验表明该算法是有效的.展开更多
介绍了基于分数阶数的变阶数最小均方算法(Fractional tap-length least mean square,FTLMS),分析算法中迭代参数对FTLMS算法收敛性能的影响。利用系统输出误差随滤波器阶数及滤波器抽头权系数迭代过程逐渐减小的特点,提出了一种新的变...介绍了基于分数阶数的变阶数最小均方算法(Fractional tap-length least mean square,FTLMS),分析算法中迭代参数对FTLMS算法收敛性能的影响。利用系统输出误差随滤波器阶数及滤波器抽头权系数迭代过程逐渐减小的特点,提出了一种新的变迭代参数的变阶数LMS算法(Variable parameter FTLMS,VP-FTLMS)。在高噪声和低噪声环境下对本文提出的算法分别进行仿真验证,并与不同的滤波器阶数迭代参数设置下的FTLMS算法相比较。仿真结果表明,与FTLMS算法相比,本文提出的VP-FTLMS算法具有更快的阶数收敛速度及更小的稳态振荡,同时获得了较小的稳态额外均方误差(Excess mean square error,EMSE)。展开更多
Parameter estimation to alpha stable distribution is difficult for without a explicit probability density function. On the base of sample characteristic function,an iterative LAD parameter estimation algorithm for SaS...Parameter estimation to alpha stable distribution is difficult for without a explicit probability density function. On the base of sample characteristic function,an iterative LAD parameter estimation algorithm for SaS is discussed. The example illustrates that the algorithm is feasible and efficient.展开更多
This paper mainly proposes a new C-XSC （C- for eXtended Scientific Computing） software for the symmetric single step method and relaxation method for computing an enclosure for the solution set and compares the meth...This paper mainly proposes a new C-XSC （C- for eXtended Scientific Computing） software for the symmetric single step method and relaxation method for computing an enclosure for the solution set and compares the methods with others＇ and then makes some modifications and finally, examples illustrating the applicability of the proposed methods are given.展开更多
文摘介绍了基于分数阶数的变阶数最小均方算法(Fractional tap-length least mean square,FTLMS),分析算法中迭代参数对FTLMS算法收敛性能的影响。利用系统输出误差随滤波器阶数及滤波器抽头权系数迭代过程逐渐减小的特点,提出了一种新的变迭代参数的变阶数LMS算法(Variable parameter FTLMS,VP-FTLMS)。在高噪声和低噪声环境下对本文提出的算法分别进行仿真验证,并与不同的滤波器阶数迭代参数设置下的FTLMS算法相比较。仿真结果表明,与FTLMS算法相比,本文提出的VP-FTLMS算法具有更快的阶数收敛速度及更小的稳态振荡,同时获得了较小的稳态额外均方误差(Excess mean square error,EMSE)。
基金Supported by Hubei Educational Committee grant Q20091809Wuhan Polytechnic University grant 2009Y21
文摘Parameter estimation to alpha stable distribution is difficult for without a explicit probability density function. On the base of sample characteristic function,an iterative LAD parameter estimation algorithm for SaS is discussed. The example illustrates that the algorithm is feasible and efficient.
文摘This paper mainly proposes a new C-XSC （C- for eXtended Scientific Computing） software for the symmetric single step method and relaxation method for computing an enclosure for the solution set and compares the methods with others＇ and then makes some modifications and finally, examples illustrating the applicability of the proposed methods are given.
