Elsevier

Neurocomputing

Volume 227, 1 March 2017, Pages 18-28
Neurocomputing

Finite-time state observer for delayed reaction-diffusion genetic regulatory networks

https://doi.org/10.1016/j.neucom.2016.09.102Get rights and content

Abstract

This paper focus on the finite-time state estimation problem for delayed reaction-diffusion genetic regulatory networks (DRDGRNs) under Dirichlet boundary conditions. The purpose is to design a finite-time state observer which is used to estimate the concentrations of mRNAs and proteins via available measurement outputs. By constructing a Lyapunov–Krasovskii functional (LKF) concluding quad-slope integrations, we establish a reaction-diffusion-dependent and delay-dependent finite-time stability criterion for the error system. The derivative of LKF is estimated by employing the Wirtinger-type integral inequality, Gronwall inequality and convex (reciprocally convex) technique. The stability criterion is to check the feasibility of a set of linear matrix inequalities (LMIs), which can be easily realized by the toolbox YALMIP of MATLAB. In addition, the expected finite-time state observer gain matrices can be represented by a feasible solution of the set of LMIs. Finally, two numerical examples are presented to illustrate the effectiveness of the theoretical results.

Introduction

In recent years, genetic regulatory networks (GRNs) has become a hot topic in many disciplines, for example, mathematics, statistics, biology and medicine, and aroused the attention of experts and scholars. As a result, a great deal of very important research results (see [1], [2], [3], [4], [5], [6], [7] and the references therein) have been achieved. GRNs, as highly complex network models, describe genetic expression and regulation behavior. The transcription and translation are the most important and most complex processes in the GRNs.

Currently, mathematical models have been one of the main tools to analyze GRNs. Due to the different forms of the GRNs and the different research purpose and methods, several GRN models have been established. For example, the Bayesian model [8], the Boolean model [7], [9] and the functional differential equations model [5], [10], [11]. A functional differential equation model describes the continuous change of mRNA and protein concentrations which have two merits: (i) the slow processes of the transcription and translation are characterized by time delays; and (ii) the continuous change of mRNA and protein concentrations are expressed as the derivatives of the unknown functions. So, functional differential equation models have been widely applied to understand the nonlinearity and complexity of GRNs. It should be emphasized that time delays are one of main sources for causing instability and/or poor performance [12], [13], [14], [15]. Accordingly, stability analysis of functional differential equation models has aroused increasing research interests, and a great number of outstanding results have been reported (see [6], [7], [16], [17], [18], [19], [20], [21], [22], [23], [24] and the references therein). The stability criteria presented in these literature are divided into two kinds: delay-dependent stability ones and delay-independent stability ones. In general, delay-dependent stability criteria are less conservative than delay-independent ones. Please refer to [25], [26], [27], [28], [29], [30], [31], [32] for effective approaches to establish delay-dependent stability criteria.

In some mathematical modeling, it is assumed that GRNs are spatially homogeneous, namely, the concentrations of mRNAs and proteins are homogenous in space at all times. However, in some cases, it is need to introduce reaction-diffusion terms into models [19], [33], [34], [35], [36], [37], [38], [39]. Especially, it is necessary to consider the diffusion of mRNAs and proteins [19], [37], [38], [39]. Thus, it is imperative to introduce reaction-diffusion terms into the continuous-time GRN models. To the best of authors' knowledge, the stability problem for delayed reaction-diffusion genetic regulatory networks (DRDGRNs) has been only studied in [19], [22], [23], [24]. Ma et al. [24] introduced DRDGRNs for the first time and established delay-dependent asymptotic stability criteria. Ma et al.'s results have been gradually improved in [19], [23] by introducing novel LKF and utilizing Wirtinger-type integral inequality approach. The problem of finite-time robust stochastic stability analysis for uncertain stochastic DRDRNs has been studied in [22].

With the change of environment, the usual feedback loops existing in GRNs may be destroyed. This will make GRNs' performance worse, and eventually lead to some fatal disease like cancer [40]. Therefore, it is necessary to adjust the feedback loops by artificial input control. For this end, the exact concentrations of the mRNAs and proteins (that is, the states of continuous GRN models) are needed. However, due to the complexity of GRNs, it is almost impossible to measure the exact concentrations. Hence, the state estimation for GRNs has been one of available methods to investigate dynamical behaviors; see, eg., [41], [42], [43], [44]. To the best knowledge of authors, the state estimation problem for DRDGRNs is only in [45], although the reaction-diffusion-free case has been researched (see [5], [41], [46], [47], [48], [49] and the references therein). This motivates our research interests.

The aim of this paper is to design a finite-time state observer for estimating concentrations of the mRNAs and proteins of DRDGRNS. A novel LKF is first constructed. Then its derivative is estimated by employing Wirtinger-type integral inequality [50], Gronwall inequality [51], convex technique and reciprocally convex technique [52]. As a result, a reaction-diffusion-dependent and delay-dependent sufficient condition is given to ensure that the error system is finite-time stable. This is different from [45] wherein the asymptotic stability of error systems are involved. The stability criterion is given in the form of linear matrix inequalities (LMIs), which can be solved by applying the Toolbox LMI or YALMIP of MATLAB. Thereby, we design a finite-time state observer whose gain matrices are described based on a feasible solution to these LMIs. In addition, two numerical examples are presented to illustrate the theoretical results obtained in this paper.

Notation: Throughout the paper, for given real symmetric matrices X and Y, X>Y(XY) means that XY is positive definite (positive semi-definite). I is the identity matrix of appropriate dimension, AT represents the transpose of matrix A. Set l={1,2,,l} for any positive integer l. Ω is a compact set in the vector space Rn with smooth boundary Ω. Let C1(X,Rn) be the Banach space of functions which map X into Rn and have the continuous first derivatives. We define a pair of norms on C1(X,Rn) and C1([d,0]×Ω,Rn) by · as follows:y(x)=(Ωy(x)Ty(x)dx)1/2andϕ(t,x)d=max{supdt0ϕ(t,x),supdt0ϕ(t,x)t,max1knsupdt0ϕ(t,x)xk},respectively. Let diag() and col() be the (block) diagonal matrix and column matrix formed by the elements in brackets, respectively.

Section snippets

Problem formulation and preliminaries

This paper considers the following DRDGRNs [19]:{m˜(t,x)t=k=1lxk(Dkm˜(t,x)xk)Am˜(t,x)+Wg(p˜(tσ(t),x))+q,p˜(t,x)t=k=1lxk(Dk*p˜(t,x)xk)Cp˜(t,x)+Bm˜(tτ(t),x),whereA=diag(a1,a2,,an),B=diag(b1,b2,,bn),C=diag(c1,c2,,cn),q=col(q1,q2,,qn),W[wij]Rn×n,Dk=diag(D1k,D2k,,Dnk),Dk*=diag(D1k*,D2k*,,Dnk*);m˜(t,x)=col(m˜1(t,x),m˜2(t,x),,m˜n(t,x)),p˜(t,x)=col(p˜1(t,x),p˜2(t,x),,p˜n(t,x)),g(p˜(t,x))=col(g(p˜1(t,x)),g(p˜2(t,x)),,g(p˜n(t,x)));x=col(x1,x2,,xl)ΩRl, Ω={x||xk|Lk,kl}, L

Observer design

In this section, we will design a state observer of the form (5) for DRDGRN (3), that is, find a pair of observer gain matrices Km and Kp so that the trivial solution of (6) is finite-time stable. For convenience, we defineei=col(0n×(i1)n,In,0n×(ni)n),i18,ς(t,x)=colem(t,x),em(tτ¯,x),em(tτ(t),x),ep(t,x),ep(tσ¯,x),ep(tσ(t),x),f^(ep(t,x)),f^(ep(tσ(t),x)),em(t,x)t,ep(t,x)t,1τ¯τ(t)tτ¯tτ(t)em(s,x)ds,1(τ¯τ(t))2tτ¯tτ(t)αtτ(t)em(s,x)dsdα,1τ(t)tτ(t)tem(s,x)ds,1τ2(t)tτ(t)tαtem(

Illustrative examples

In this section, we will give two numerical examples to demonstrate the availability of the state observer proposed previously.

Example 1

Consider DRDGRN (3) with measurements (4), whereA=diag(0.2,1.1,1.2),B=diag(1.0,0.4,0.7),C=diag(0.3,0.7,1.3),L1=L2=L3=1,W=[000.50.50000.50],D1=D2=D3=diag(0.1,0.1,0.1),D1*=D2*=D3*=diag(0.2,0.2,0.2),M=[0.50.600.30.80.2],N=[0.70.250.30.40.20.3],fi(x)=x21+x2,i3. When μτ=μσ=1.5,τ¯=σ¯=1.1, c1=1, c2=5, T=10 and α=0.002, by using the Toolbox YALMIP of MATLAB, we find

Conclusions

This paper investigates the finite-time state estimation problem for a class of DRDGRNs under Dirichlet boundary conditions. A finite-time state observer has been designed to estimate the concentrations of mRNAs and proteins based on available network outputs. A LKF including new integral terms is first introduced. Then its derivative is estamated by using the so-called Green's second identity, Wirtinger-type inequality, Gronwall inequality, convex technique and reciprocally convex technique.

Acknowledgements

This work was supported in part by the National Natural Science Foundation of China (11371006), the Natural Science Foundation of Heilongjiang Province (F201326 and A201416), the Fund of Heilongjiang Education Committee (12541603), and the Heilongjiang University Innovation Fund for Graduates (YJSCX2016-027HLJU).

The authors thank the anonymous associate editor and referees for their helpful comments and suggestions.

X.F. Fan was born in Inner Mongolia, China, in 1991. She received the B.S. degree in School of Mathematical Science from Hulunbeir College, Hulunbeir, China, in 2014. She is currently an M.S. student in Heilongjiang University, Harbin, China. Her research interests include genetic regulatory networks and stability analysis of delayed dynamic systems.

References (56)

  • J.P. Zhou et al.

    Finite-time robust stochastic stability of uncertain stochastic delayed reaction-diffusion genetic regulatory networks

    Neurocomputing

    (2011)
  • H.B. Zeng et al.

    New results on stability analysis for systems with discrete distributed delay

    Automatica

    (2015)
  • J.L. Wang et al.

    The stability analysis of a general viral infection model with distributed delays and multi-staged infected progression

    Commun. Nonlinear Sci. Numer. Simul.

    (2015)
  • P. Shi et al.

    Stochastic finite-time state estimation for discrete time-delay neural networks with Markovian jumps

    Neurocomputing

    (2015)
  • J.L. Wang et al.

    Passivity-based synchronization of a class of complex dynamical networks with time-varying delay

    Automatica

    (2015)
  • R.W. Johnstone et al.

    Apoptosis: a link between cancer genetics and chemotherapy

    Cell

    (2002)
  • R. Anbuvithya et al.

    Sampled-data state estimation for genetic regulatory networks with time-varying delays

    Neurocomputing

    (2015)
  • L.-P. Tian et al.

    State observer design for delayed genetic regulatory networks

    Comput. Math. Method. M

    (2014)
  • V. Vembarasan et al.

    State estimation for delayed genetic regulatory networks based on passivity theory

    Math. Biosci.

    (2013)
  • M. Mohammadian et al.

    An LPV based robust peak-to-peak state estimation for genetic regulatory networks with time varying delay

    Neurocomputing

    (2015)
  • X.B. Wan et al.

    Robust non-fragile H state estimation for discrete-time genetic regulatory networks with Markov jump delays and uncertain transition probabilities

    Neurocomputing

    (2015)
  • Y.T. Wang et al.

    H filtering for discrete-time genetic regulatory networks with random delay described by a Markovian chain

    Abstr. Appl. Anal.

    (2014)
  • P.G. Park et al.

    Reciprocally convex approach to stability of systems with time-varying delays

    Automatica

    (2011)
  • P. Park et al.

    Auxiliary function-based integral inequalities for quadratic functions and their applications to time-delay systems

    J. Frankl. Inst.

    (2015)
  • A. Seuret et al.

    Wirtinger-based integral inequality: application to time-delay systems

    Automatica

    (2013)
  • J.Q. Hu et al.

    Stabilization of genetic regulatory networks with mixed time-delays: an adaptive control approach

    IMA J. Math. Control Inf.

    (2015)
  • X. Zhang et al.

    Globally asymptotic stability analysis for genetic regulatory networks with mixed delays: an M-matrix-based approach

    IEEE/ACM Trans. Comput. Biol. Bioinform.

    (2016)
  • M.J. Beal et al.

    A Bayesian approach to reconstructing genetic regulatory networks with hidden factors

    Bioinformatics

    (2005)
  • Cited by (28)

    • Novel adaptive strategies for synchronization control mechanism in nonlinear dynamic fuzzy modeling of fractional-order genetic regulatory networks

      2022, Chaos, Solitons and Fractals
      Citation Excerpt :

      GRNs can be naturally modeled as dynamical systems, which are powerful tools for gaining biological insights based on their complex dynamical behaviors [2,3]. In the past decades, many researchers have made intensive efforts to understand the dynamical properties of genetic regulation inside cells, including system modeling, stability analysis, filtering and so on (see [4–6] and the references therein). Furthermore, the GRNs, due to the slow processes of the transcription and translation, the mRNA and protein concentrations not only depend on the current state, but also rely on the delayed states.

    • Robust stability of delayed Markovian switching genetic regulatory networks with reaction–diffusion terms

      2020, Computers and Mathematics with Applications
      Citation Excerpt :

      In living organism, there are a great number of genes and proteins directly or indirectly interact with each other, which make up of a dynamic GRNs acts as a complex dynamical system. The mathematical models of GRNs are helpful to investigate gene regulation progress such as Bayesian model [1–3], Boolean model [4–8] and differential genetic networks [9–18]. Among these models, the differential equation model is a kind of continuous system which is able to provide more accurate and detailed depiction of dynamic behavior.

    View all citing articles on Scopus

    X.F. Fan was born in Inner Mongolia, China, in 1991. She received the B.S. degree in School of Mathematical Science from Hulunbeir College, Hulunbeir, China, in 2014. She is currently an M.S. student in Heilongjiang University, Harbin, China. Her research interests include genetic regulatory networks and stability analysis of delayed dynamic systems.

    Y. Xue received the B.S. degree in Automation from Harbin University of Science and Technology, China in 2002; the Ph.D degree in Control Theory and Control Engineering from Harbin Institute of Technology, China in 2007. From April 2007 to August 2009, she is an Engineer in Sichuan Academy of Aerospace Technology. From September 2009 to May 2013, she is a Senior Engineer in Sichuan Academy of Aerospace Technology. In June 2013, she joined Heilongjiang University. Her research interests include stability analysis of delayed dynamic systems, robust control and genetic regulatory networks.

    X. Zhang received Ph.D. degree in Control Theory from Queen's University of Belfast in UK in 2004. Since 2004 he has been at Heilongjiang University, where he is currently a Professor in the School of Mathematical Science. His current research interests include neural networks, genetic regulatory networks, mathematical biology and stability analysis of delayed dynamic systems. He has received the Second Class of Science and Technology Awards of Heilongjiang Province. He is a member of the IEEE, and a Vice President of Mathematical Society of Heilongjiang Province. Since 2006, he has been serving as an Editor of the Journal of Natural Science of Heilongjiang University. He has authored more than 100 research papers.

    J. Ma was born in Heilongjiang, China, in 1979. She received the M.E. and Ph.D. degrees from Heilongjiang University, Harbin, China, in 2007 and 2012, respectively. Her current research interests include the networked systems filtering and information fusion filtering.

    View full text