基于递归神经网络的基因调控网络稳定性分析  

Stability Analysis of Genetic Regulatory Networks Based on Recurrent Neural Networks

作　　者：朱延正[1,2] 李玉榕[1,2] 杜民 

机构地区：[1]福州大学电气工程与自动化学院,福州350108 [2]福建省医疗器械与医药技术重点实验室,福州350002

出　　处：《系统仿真学报》2012年第12期2506-2510,共5页Journal of System Simulation

基　　金：科技部国际合作项目(2009DFA32050);福建省自然科学基金(2009J01280)

摘　　要：基因调控网络是一类复杂的非线性动力学系统,其稳定性分析在生命科学研究中发挥着重要作用。基于递归神经网络所具有的较强的建模能力,建立了一类新的时滞基因调控网络模型。根据模型的特点构建合适的Lyapunov-Krasovskii函数,并借助线性矩阵不等式技术与S-方法推导出使得网络全局渐进与指数稳定的充分性条件,最后通过实例验证了此方法的有效性。Genetic regulatory networks is a kind of complex nonlinear dynamics system. The stability analysis of which plays an important role in the study of life science. One novel genetic regulatory networks was defined with the help of strong modeling ability of recurrent neural networks. By constructing an appropriate Lyapunov-Krasovskii function according to the characteristics of this developed model, the global asymptotic and exponential stability were established through linear matrix inequality techniques and S-approach. Finally, the performance and effectiveness of the proposed method were illustrated with numerical examples.

关 键 词：递归神经网络 标准基因调控网络 时变时滞 稳定性 

分 类 号：TP11[自动化与计算机技术—控制理论与控制工程]