Developing Objective Sensitivity Analysis of PeriodicSystems: Case Studies of Bi
作者：LU Bao-Yun1,2 YUE Hong1,3 单位：1. Key Laboratory of Complex Systems and Intelligence Sci-ence,Institute of Automation,Chinese Academy of Sciences,Beijing100190,P. R. China 2. College of Computer Science and Technol-ogy,Henan Polytechnic University,Jiaozuo 454000,P. R. China 3.Departmen 本文刊于：
Sensitivity analysis periodic systems ph
Sensitivity analysis is a powerful tool in investigating the impact of parameter variations on the change of systembehaviours quantitatively. For a periodic system, sensitivity analysis is a challenging problem since the standard sensitivity metricsgrow unbounded when time tends to infinity. Objective sensitivity analyses using various oscillation features such as period, phase,amplitude, etc. are therefore needed to circumvent this problem. In this work, a new concept of basal state sensitivity is proposedbased on which a phase sensitivity calculation is derived. The improved period sensitivity calculation following an existing algorithmusing singular value decomposition (SVD) is also presented, which provides a simple calculation for the basal state sensitivity. Thesenew sensitivity calculations are developed with the purpose to analyse biological oscillators since there is an increasing interest inunderstanding how oscillations occur and what the main controlling factors are following a growing experimental and computationalevidence of oscillations in biological systems. The improved calculation of period sensitivity is shown to be consistent with theprevious methods through a well studied circadian rhythm model. The calculation of new objective sensitivities are also testified bythe same circadian rhythm model as well as an oscillatory signal transduction pathway model, which further illustrates the efficiencyof this approach in handling complex biological oscillators in the presence of reaction conservations.