基于动态优化算法的复杂生物网络的状态调控  

State regulation for complex biological networks based on dynamic optimization algorithms

作　　者：揭豪 袁美晨 朱国柱 洪伟荣[1] JIE Hao;YUAN Meichen;ZHU Guozhu;HONG Weirong(College of Energy Engineering,Zhejiang University,Hangzhou 310027,P.R.China)

机构地区：[1]浙江大学能源工程学院

出　　处：《生物医学工程学杂志》2020年第1期19-26,共8页Journal of Biomedical Engineering

摘　　要：近年研究表明,某些药物能改变基因调控网络(GRNs)中的调控反应参数,促使细胞从病变状态转移到正常状态。在此基础上,有学者构建了基于吸引子和分岔理论的状态调控框架。现有的控制策略中,控制信号规律都是自行拟定的,并且所使用的参数扰动法仅能求出单控制变量情况下状态转移所需的时间。因此,针对以非线性常微分方程组(ODEs)形式描述的复杂生物网络,本文提出了一种基于动态优化算法改变网络状态的最优调控方法,在掌握生物网络基本特性的基础上建立了动态优化问题,并以简单低维的三节点GRN和复杂高维的癌症GRN为例,利用MATLAB计算得到单控制变量及多控制变量条件下的最优控制策略。本文所提方法旨在实现精确且快速的状态调控,为试验研究和实际治疗提供参考依据。Recent studies showed that certain drugs can change regulatory reaction parameters in gene regulatory networks(GRNs)and therefore restore pathological cells to a normal state.A state control framework for regulating biological networks has been built based on attractors and bifurcation theory to analyze this phenomenon.However,the control signal is self-developed in this framework,of which the parameter perturbation method can only calculate the state transition time of cells with single control variable.Therefore,an optimal control method based on the dynamic optimization algorithms is proposed for complex biological networks modeled by nonlinear ordinary differential equations(ODEs).In this approach,dynamic optimization problems are constructed based on basic characteristics of the biological networks.Furthermore,using an example of a simple low-dimensional three-node GRN and a complex highdimensional cancer GRN,MATLAB is utilized to calculate optimal control strategies with either single or multiple control variables.This method aims to achieve accurate and rapid state regulation for biological networks,which can provide a reference for experimental researches and medical treatment.

关 键 词：基因调控网络 状态调控 动态优化 最优控制策略 

分 类 号：Q811.4[生物学—生物工程] TP18[自动化与计算机技术—控制理论与控制工程] O175.14[自动化与计算机技术—控制科学与工程]