Parameters Identification by a Piecewise Adaptive Rule with a Fractional Power
Parameters Identification by a Piecewise Adaptive Rule with a Fractional Power
复制标题
DOI：
10.1142/s0218127415501667
复制
发表时间：
2015-12
期刊：
影响因子：
--
通讯作者：
Bing Xu;Wei Lin
中科院分区：
文献类型：
--
作者：
Bing Xu;Wei Lin
解开真实的系统（如物理系统和生物系统）中的参数，是模型驱动研究（包括计算系统生物学）的中心议题之一。现有的自适应同步技术虽然在理论上被证明是有效的参数识别在一些代表性的模型，总是表现出不令人满意的收敛速度时，参数的数量增加，并考虑到真实的数据。本文提出了一种基于有限时间控制的分数阶分段自适应律，以加快参数辨识的速度。从理论上证明了该规则的全局可行性，并通过几个具有物理和生物学意义的例子加以说明。
Unraveling parameters in real systems, such as physical systems and biological systems, is one of the central topics in model-driven research including computational systems biology. The existing adaptive synchronization technique, though theoretically proved to be valid for parameters identification in some representative models, is always showing unsatisfactory convergence rate when the number of parameters increases and real data are taken into account. In this paper, a piecewise adaptive rule with a fractional power based on finite time control is proposed to accelerate the rate for parameter identification. The global feasibility of the proposed rule is proved theoretically and illustrated by several examples of physical and biological significance.