Asymptotic stability of cellular neural networks with multiple proportional delays☆
Introduction
Cellular neural networks (CNNs) [1] have been widely investigated owing to their widespread applications in image processing, pattern recognition, optimization and associative memories. These practical application of CNNs depends on the existence and stability of the equilibrium point of the CNNs. Moreover, time delay is inevitable due to the finite switching speed of information processing and the inherent communication time of neurons, and its existence may cause the instability of the system. Therefore, a number of stability criteria of CNNs with delays have been proposed [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20]. At present, stability results of CNNs with delays studied in [4], [6], [11], [13], [16], [18] are mainly based on such approaches as -matrix, algebraic inequalities, and so on. As pointed out in [14], the characteristic of those results, which means to take absolute value operation on the interconnection matrix, leads to the ignorance of neuron’s inhibitory and excitatory effects on neural networks. In recent years, linear matrix inequality (LMI) technique has been used to deal with the stability problem for neural networks [2], [3], [5], [7], [8], [9], [10], [12], [14], [15], [17], [19], [20]. The feature of LMI-based results is that it can consider the neuron’s inhibitory and excitatory effects on neural networks. However, few stability results have been obtained for CNNs with proportional delays on basis of LMI, which is important, as did for neural networks model studied in [32], [33]. The exponential stability of CNNs with multi-pantograph delays was studied by nonlinear measure in [32]. In [33], by employing matrix theory and Lyapunov functional, global exponential stability of a class of CNNs with multi-proportional delays was investigated and delay-dependent sufficient conditions were obtained. The proportional delay system as an important mathematical model often rises in some fields such as physics, biology systems and control theory and it has attracted many scholars’ interest [25], [26], [27], [28], [29], [30], [31].
As a neural network usually has a spatial nature for the presence of an amount of parallel pathways of a variety of axon sizes and lengths, it is desired to model by introducing continuously proportional delay over a certain duration of time. The proportional delay function is that as (), so it is time-varying, unbounded, and monotonically increasing. Therefore, the network’s running time may be controlled according to the network allowed delays. At the present time, besides the distributed delays, the delay function is usually required to be bounded in the stability discussion of neural networks with delays, such as, these results in [9], [12], [19], [20], [21], [22], [23], [24], are required that the delay function satisfies and other conditions. It is seldom that the delay function is unbounded, i.e. . Compared with the distributed delay [6], [7], [8], [10], [13], whose delay kernel functions satisfy some conditions such that the distributed delay is easier to be handled [8], [10], [13] in the use inequality, but the proportional delay is not easy to be controlled. Thus, it is important to study stability of neural networks with proportional delays in theory and practice.
Motivated by the discussion above, the asymptotic stability of CNNs with multiple proportional delays is further discussed in this paper, inspired by Liao et al. [4], Zhang and Wang [14] and Zhang and Zhou [32]. This paper is organized as follows. Model description and preliminaries are given in Section 2. Two delay-dependent and delay-independent sufficient conditions are given in Section 3 to ascertain the asymptotic stability of the CNNs with multiple proportional delays, which are easy to be verified. Numerical examples and their simulation are presented in Section 4 to illustrate the effectiveness and less conservative of obtained results. Finally, conclusions are given in Section 5.
Section snippets
Model description and preliminaries
Consider the following CNNs with multiple proportional delaysfor , where n is the number of neurons; is a constant; is the state variable; and are constants which denote the strengths of connectivity between the cells j and i at time t and , respectively; denotes a nonlinear activation function; denotes the constant external inputs; are proportional delay
Main results
Theorem 3.1 The origin of system (2.3) is globally asymptotically stable, if there exist positive diagonal matrices and a constant , such that the following inequality holds:where is an square matrix, whose ith row is composed of and other rows are all zeros, and . Proof Consider the following Lyapunov
Numerical examples
Example 4.1 Consider the neural network model (2.1), wherethe activation function is described by with . Clearly, satisfy condition (2.2) above. By some simple calculations, we obtain , and . And Take and . Applying Theorem 3.1, we have
Conclusions
New sufficient conditions are derived for asymptotic stability of equilibrium point for cellular neural networks with multiple proportional delays, which is different from the existing ones and has wider application fields. It can be shown that the derived criteria are less conservative than previously existing results through the numerical example. The obtained results can be expressed in the form of linear matrix inequality and be easy to be verified by utilizing Matlab LMI toolbox. These
References (36)
- et al.
Delay-depend exponential stability of cellular neural networks with time-varying delays
Chaos Solitons Fractals
(2005) - et al.
Global exponential periodicity and stability of cellular neural networks with variable and distributed delays
Appl. Math. Comput.
(2008) - et al.
Global exponential stability of cellular neural networks with time-varying discrete and distributed delays
Neurocomputing
(2009) - et al.
Global asymptotic stability of stochastic fuzzy cellular neural networks with multi time-varying delays
Expert Syst. Appl.
(2010) - et al.
Novel stability of cellular neural networks with interval time-varying delay
Neural Networks
(2008) - et al.
Almost sure exponential stability of stochastic cellular neural networks with unbounded distributed delays
Neurcomputing
(2009) The asymptotic behavior of certain difference equation with proportional delays
Comput. Math. Appl.
(1994)On neural functional-differential equation with proportional delays
J. Math. Anal. Appl.
(1997)- et al.
Cellular neural networks: theory and applications
IEEE Trans. Circuits Syst. I
(1988) Global asymptotic stability of cellular neural networks with unequal delays: LMI approach
Electron. Lett.
(2004)
Global asymptotic stability and global exponential stability of delayed cellular neural networks
IEEE Trans. Circuits Syst. II
An improved global asymptotic stability criterion for delayed cellular neural networks
IEEE Trans. Neural Networks
Global asympotic stability of fuzzy cellular neural networks with unbounded distributed delays
Neural Process. Lett.
Stability and dissipativity analysis of distributed delay cellular neural networks
IEEE Trans. Neural Networks
Global exponential stability analysis for cellular neural networks with variable coefficients and delays
Neural Comput. Appl.
Global asymptotic stability of delayed cellular neural networks
IEEE Trans. Neural Networks
Global exponential stability for uncertain cellular neural networks with time-varying delays via LMI approach
Chaos Solitons Fractals
Equilibrium analysis for improved signal range model of delayed cellular neural networks
Neural Process. Lett.
Cited by (51)
Fixed-time sliding mode controller design for a class of nonlinear Markovian jumping systems
2020, Journal of the Franklin InstituteCitation Excerpt :In [32], the global asymptotic stability of equilibria in autonomous systems of differential equations are analyzed. In [33], the asymptotic stability of the equilibrium point of cellular neural networks with multiple proportional delays is presented. In order to improve the convergence rate of closed-loop system, the concept of fixed-time stabilization (FTS) [34,35] is introduced in this paper.
- ☆
This work is supported by the National Science Foundation of China (Nos. 61374009, 61272514, 61003287, 61170272, 61121061, 61161140320), Tianjin Municipal Education Commission (No. 20100813), NCET (No. NCET-13-0681), the Specialized Research Fund for the Doctoral Program of Higher Education (No. 20100005120002), the Fok Ying Tong Education Foundation (No. 131067) and the Fundamental Research Funds for the Central Universities (No. BUPT2012RC0221).