Table of Contents Author Guidelines Submit a Manuscript
Journal of Applied Mathematics
Volume 2014 (2014), Article ID 468584, 10 pages
http://dx.doi.org/10.1155/2014/468584
Research Article

Stability and Hopf Bifurcation of an n-Neuron Cohen-Grossberg Neural Network with Time Delays

1Institute of Applied Mathematics, Shijiazhuang Mechanical Engineering College, Shijiazhuang 050003, China
2Department of Information Engineering, Shijiazhuang Mechanical Engineering College, Shijiazhuang 050003, China

Received 9 October 2013; Revised 28 January 2014; Accepted 29 January 2014; Published 6 March 2014

Academic Editor: Wan-Tong Li

Copyright © 2014 Qiming Liu and Sumin Yang. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

A Cohen-Grossberg neural network with discrete delays is investigated in this paper. Sufficient conditions for the existence of local Hopf bifurcation are obtained by analyzing the distribution of roots of characteristic equation. Moreover, the direction and stability of Hopf bifurcation are obtained by applying the normal form theory and the center manifold theorem. Numerical simulations are given to illustrate the obtained results.

1. Introduction

In recent years, more and more mathematicians, biologists, physicists, and computer scientists focus on artificial neural networks. It is well known that the analysis of the dynamical behaviors is a necessary step for practical design of neural networks since their applications heavily depend on the dynamical behaviors; many important results on dynamical behaviors of neural networks have been obtained [123]. The neural networks are large-scale and complex systems, and the dynamical behaviors of neural networks with delays are more complicated; in order to obtain a deep and clear understanding of the dynamics of complicated neural networks with time delays, researchers have focused on the studying of simple systems [1222]. This is indeed very useful since the complexity found may be carried over to large neural networks.

The research on dynamical behaviors of neural networks involves not only the dynamic analysis of equilibrium but also that of periodic solution, bifurcation, and chaos; especially, the periodic oscillatory behavior of the neural networks is of great interest in many applications [2, 3]. Since periodic oscillatory can arise through the Hopf bifurcation in different system with or without time delays, it is very important to discuss the Hopf bifurcation of neural networks.

In 1983, Cohen-Grossberg [1] proposed a kind of neural networks, which are now called Cohen-Grossberg neural networks. The networks have been successfully applied to signal processing, pattern recognition, optimization, and associative memories. Recently, some results on the existence and globally asymptotical stability of periodic Cohen-Grossberg neural networks have been obtained [715]. However, up to now, to the best of the author’s knowledge, bifurcation of Hopfield neural networks has been discussed by many researchers [1219], but few results on the bifurcation of Cohen-Grossberg neural networks have been obtained. Zhao discussed the bifurcation of a two-neuron discrete-time Cohen-Grossberg neural network in [20] and the bifurcation of a two-neuron continuous-time Cohen-Grossberg neural network with distributed delays in which kernel function is in [21]. We discussed the bifurcation of a two-neuron Cohen-Grossberg neural network with discrete delays in [22]. The objective of this paper is to study the following -neuron continuous-time Cohen-Grossberg neural network with discrete delays and ring architecture: where denote the state variable of the th neuron; represent amplification functions which are positive for ; denote the signal functions of the th neuron; are appropriately behaved functions; are connection weights of the neural networks; discrete delays correspond to the finite speed of the axonal signal transmission: , .

Ring architectures have been found in variety of neural structures, and they are investigated to gain insight into the mechanisms underlying the behaviors of recurrent neural networks [23].

The rest of this paper is organized as follows. Stability property and existence of Hopf bifurcation for system (1) are obtained in Section 2. Based on the normal form method and the center manifold, the formulas for the direction of Hopf bifurcation and stability of the bifurcating periodic solutions are derived in Section 3. An example is given in Section 4 to illustrate the main results, and conclusions are drawn in Section 5.

2. Stability Analysis and Existence of Local Bifurcation

Lemma 1 (see [11]). Consider the exponential polynomial where and are constants. Then as vary, the sum of the order of zeros of on the open right half plane can change only if a zero appears on or across the imaginary axis.

In the following discussion, for convenience, we denote

Throughout this paper, we assume that(H1), , and ;(H2)there exist constants, such thatfor.

From assumption (H1)-(H2) that the origin is an equilibrium of system (1).

Let

System (1) can be transformed into the following equivalent system: where .

The linear system of system (5) around the equilibrium is given by

The associated characteristic equation of system (5) is

Suppose that is a root of the characteristic equation, where is imaginary unit which satisfies . Substituting into (7), then we have

Separating the real and imaginary parts of (8), we have where in which denotes the inverse of the cotangent function.

Since for and

Hence is an increasing bijective function.

We know from the second equation in (9) that . Denote , ; then we have from the first equation in (9) that

When .

Furthermore, from the value given above, we have

Obviously, if is a root of equation (7), is also the root of equation (7). This implies that is a pair of purely imaginary roots of equation (7). On the other hand, we have from (7) that where

Hence we have

Since the roots of the characteristic equation (7) are , when , so the equilibrium of system (5) is asymptotically stable. As the parameter varies on the open right half plane can change only if a zero appears on or across the imaginary axis. According to Lemma 1 and (14), we obtain that the equilibrium of system (5) is asymptotically stable if and only if .

When , the characteristic equation of system (5) has a simple root , and all the other roots have negative real parts. A pitchfork bifurcation may occur at the origin in system (5) [16].

When , , the characteristic equation of system (3) has a pair of purely imaginary roots , and all the other roots have negative real parts. Note that due to according to (11), so, ; that is, . We also know from (17) that . System (3)undergoes a Hopf bifurcation which occurs at the origin when .

From the above discusses, Lemma 1, and the Hopf bifurcation theorem in [24] for functional differential equations, we have the following results.

Theorem 2. Under assumptions (H1)-(H2), we have the following:(1)if , the equilibrium of system (1) is asymptotically stable;(2)if , the equilibrium of system (1) is unstable;(3)if , a Hopf bifurcation occurs at the origin in system (1),where in which and satisfies the equation , .

3. Direction and Stability of Hopf Bifurcation for the Network

In this section, we will derive explicit formulas for determining the properties of the Hopf bifurcation at critical by using the normal form theory and the center manifold theorem [25], and we always make , vary with a parameter and the other ones are fixed.

We still discuss system (5). For the sake of generality, let . Denote and , dropping the bars for simplification of notation; then system (5) can be written as functional differential equation in as where , , and are given, respectively, by where where in which

From the discussions in Section 2, we know that if , system (19) undergoes a Hopf bifurcation at the equilibrium , and the associated characteristic equation of system (19) has a pair simple imaginary roots .

By the Riesz representation theorem, there exists a bounded variation function for such that

In fact, we can choose where is the Dirac delta function and

For , define

The system (19) can be transformed into the following operator equation form: where for .

Denote

For , define

For and , we define a bilinear form

Then and are adjoint operators. We know that are eigenvalues of , so are also eigenvalues of .

Now we compute the eigenvectors of and corresponding to and .

Suppose that is the eigenvector of corresponding to ; then . From (25) and (26), we know

Similarly, we know that with and is the eigenvector of corresponding to , where

Moreover, and .

Using the same notations as Hassard et al. [25], we construct the coordinates to describe the center manifold at .

Define

On the center manifold , we have where

and are local coordinates for the center manifoldin the direction of and . Note that is real if is real. We only consider real solutions.

For solution of (19), since , we have

We rewrite this as with

From (29), (36), and (39), we have where

Since where in which ,  .

Denote the th element of by and the th element of by

Then if follows that

Substitute (46) into (43) and comparing the coefficients in (40) with those in (43), we have where is th element of shown in (32).

Since there are and in , we still need to figure them out. Note that on the center manifold , we have

We have from (41), (42), and (48) that

Equations (39) and (41) mean for .

Comparing the coefficients with (42), we have

From (26), (49), and (50), we can obtain

So and similarly where , , , , , .

In the following, we focus on the computation of and . From (26) and (49), we know in which

We know from (41) and (43) that in which in which

Substituting (53) and (57) into (55), we have

Solving this we can obtain .

Similarly, we can obtain from

Based on the above analysis, we can see that each in (47) is determined by the parameters and delays for (1). Thus, we can compute the following quantities:

It is known that determines the direction of the Hopf bifurcation and determines the stability of the bifurcating periodic solutions. Since , we know if ; then the Hopf Bifurcation is supercritical (subcritical), the bifurcating periodic solutions exist for , and the bifurcating periodic solutions are stable (unstable). determines the period of the bifurcating periodic solutions: the period increases (decreases) if .

Under some conditions, the equilibrium of system (1) is globally asymptotically stable. The following result can be directly obtained from Corollary  2 in [5].

Theorem 3. Under assumptions (H1)–(H3), the equilibrium of system (1) is globally asymptotically stable if the following conditions hold.(H3)There exist constants such that for .(H4)There exist positive constants such that for .(H5)The following matrix is an -matrix:

Note that the conditions in Theorem 3 have more restrictions than those in Theorem 2. Since is an M-matrix, we have [26]; that is, ; it yields , which, together with conditions (H3) and (H4), implies that ; moreover, due to . Hence, conditions (H3)–(H5) imply that the condition in Theorem 2 holds.

4. A Numerical Example

Example 1. Consider the following Cohen-Grossberg neural network with discrete delays:
We can obtain that and furthermore we obtain that in view of bisection method by using MATLAB. It is easy to know. We also know from (3) that .

According to Theorem 2, the zero solution of system (64) is asymptotically stable when , and when , the Hopf bifurcation occurs at the origin.

Case 1. Let . ; then the zero solution of system (64) is asymptotically stable. Figure 1 shows the dynamic behaviors of system (64) with initial condition .

468584.fig.001
Figure 1: Phase plot in space for system (64) with .

Case 2. Let , and . We know from Theorem 2 that the Hopf bifurcation occur at the origin; furthermore, we can obtain , so the bifurcating periodic solutions are supercritical and asymptotically stable. Figures 2 and 3 show the dynamic behaviors of system (64) with initial conditions and , respectively.

468584.fig.002
Figure 2: Phase plot in space for system (64) with .
468584.fig.003
Figure 3: Phase plot in space for system (64) with .

The presented numerical simulations illustrate the theoretical results.

5. Conclusions

An -neuron Cohen-Grossberg neural network with discrete delays and ring architecture is analyzed in this paper. By using as a bifurcation parameter, we show that this system undergoes a Hopf bifurcations at a critical parameter: where and satisfies the equation , . The direction of Hopf bifurcation and the stability of the bifurcating periodic solutions are investigated by applying the normal form theory and the center manifold theorem for continuous time system. The phenomena of bifurcating periodic solutions for Cohen-Grossberg neural networks coincide with the fact that learning usually requires repetition [2], and periodic sequences of neural impulse are also of fundamental significance for the control of dynamic functions of the body such as heart beat which occurs with great regularity and breathing [19]. In this paper, we extend the results about the existence of local Hopf bifurcation in [22] to the case of a discrete-time -neuron Cohen-Grossberg system with discrete delays. In the future, the problem for the existence of global Hopf bifurcation will be expected to be solved.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

The author is greatly indebted to the reviewers and the editors for their very valuable suggestions and comments which improved the quality of the presentation. This research was supported by the Hebei Provincial Natural Science Foundation of China under Grant no. A2012205028 and the Innovation Foundation of Shijiazhuang Mechanical Engineering College under Grant no. Yscx1201.

References

  1. M. A. Cohen and S. Grossberg, “Absolute stability of global pattern formation and parallel memory storage by competitive neural networks,” IEEE Transactions on Systems, Man, and Cybernetics, vol. 13, no. 5, pp. 815–826, 1983. View at Publisher · View at Google Scholar · View at MathSciNet
  2. S. Townley, A. Ilchmann, M. G. Weiß et al., “Existence and learning of oscillations in recurrent neural networks,” IEEE Transactions on Neural Networks, vol. 11, no. 1, pp. 205–214, 2000. View at Publisher · View at Google Scholar · View at Scopus
  3. Z. Huang and Y. Xia, “Exponential periodic attractor of impulsive BAM networks with finite distributed delays,” Chaos, Solitons & Fractals, vol. 39, no. 1, pp. 373–384, 2009. View at Publisher · View at Google Scholar · View at MathSciNet
  4. Z. Guo and L. Huang, “LMI conditions for global robust stability of delayed neural networks with discontinuous neuron activations,” Applied Mathematics and Computation, vol. 215, no. 3, pp. 889–900, 2009. View at Publisher · View at Google Scholar · View at MathSciNet
  5. Q. Song and J. Zhang, “Global exponential stability of impulsive Cohen-Grossberg neural network with time-varying delays,” Nonlinear Analysis: Real World Applications, vol. 9, no. 2, pp. 500–510, 2008. View at Publisher · View at Google Scholar · View at MathSciNet
  6. C. Bai, “Stability analysis of Cohen-Grossberg BAM neural networks with delays and impulses,” Chaos, Solitons & Fractals, vol. 35, no. 2, pp. 263–267, 2008. View at Publisher · View at Google Scholar · View at MathSciNet
  7. W. Lu and T. Chen, “R+n-global stability of a Cohen-Grossberg neural network system with nonnegative equilibria,” Neural Networks, vol. 20, no. 6, pp. 714–722, 2007. View at Publisher · View at Google Scholar · View at Scopus
  8. X. Li, “Existence and global exponential stability of periodic solution for impulsive Cohen-Grossberg-type BAM neural networks with continuously distributed delays,” Applied Mathematics and Computation, vol. 215, no. 1, pp. 292–307, 2009. View at Publisher · View at Google Scholar · View at MathSciNet
  9. Y. Li, X. Chen, and L. Zhao, “Stability and existence of periodic solutions to delayed Cohen-Grossberg BAM neural networks with impulses on time scales,” Neurocomputing, vol. 72, no. 7-9, pp. 1621–1630, 2009. View at Publisher · View at Google Scholar · View at Scopus
  10. H. Xiang and J. Cao, “Exponential stability of periodic solution to Cohen-Grossberg-type BAM networks with time-varying delays,” Neurocomputing, vol. 72, no. 7-9, pp. 1702–1711, 2009. View at Publisher · View at Google Scholar · View at Scopus
  11. S. Ruan and J. Wei, “On the zeros of transcendental functions with applications to stability of delay differential equations with two delays,” Dynamics of Continuous, Discrete & Impulsive Systems. Series A, vol. 10, no. 6, pp. 863–874, 2003. View at Google Scholar · View at MathSciNet
  12. J. Wei and S. Ruan, “Stability and bifurcation in a neural network model with two delays,” Physica D, vol. 130, no. 3-4, pp. 255–272, 1999. View at Publisher · View at Google Scholar · View at MathSciNet
  13. J. Cao and M. Xiao, “Stability and Hopf bifurcation in a simplified BAM neural network with two time delays,” IEEE Transactions on Neural Networks, vol. 18, no. 2, pp. 416–430, 2007. View at Publisher · View at Google Scholar · View at Scopus
  14. S. Guo, L. Huang, and L. Wang, “Linear stability and Hopf bifurcation in a two-neuron network with three delays,” International Journal of Bifurcation and Chaos, vol. 14, no. 8, pp. 2799–2810, 2004. View at Publisher · View at Google Scholar · View at MathSciNet
  15. C. Huang, L. Huang, J. Feng, M. Nai, and Y. He, “Hopf bifurcation analysis of a two-neuron network with four delays,” Chaos, Solitons & Fractals, vol. 34, no. 3, pp. 795–812, 2007. View at Publisher · View at Google Scholar · View at MathSciNet
  16. J. Wei and C. Zhang, “Bifurcation analysis of a class of neural networks with delays,” Nonlinear Analysis: Real World Applications, vol. 9, no. 5, pp. 2234–2252, 2008. View at Publisher · View at Google Scholar · View at MathSciNet
  17. X. Zhou, Y. Wu, Y. Li, and X. Yao, “Stability and Hopf bifurcation analysis on a two-neuron network with discrete and distributed delays,” Chaos, Solitons & Fractals, vol. 40, no. 3, pp. 1493–1505, 2009. View at Publisher · View at Google Scholar · View at MathSciNet
  18. Y. Yang and J. Ye, “Stability and bifurcation in a simplified five-neuron BAM neural network with delays,” Chaos, Solitons & Fractals, vol. 42, no. 4, pp. 2357–2363, 2009. View at Publisher · View at Google Scholar · View at MathSciNet
  19. Y. Song, M. Han, and J. Wei, “Stability and Hopf bifurcation analysis on a simplified BAM neural network with delays,” Physica D, vol. 200, no. 3-4, pp. 185–204, 2005. View at Publisher · View at Google Scholar · View at MathSciNet
  20. H. Zhao and L. Wang, “Stability and bifurcation for discrete-time Cohen-Grossberg neural network,” Applied Mathematics and Computation, vol. 179, no. 2, pp. 787–798, 2006. View at Publisher · View at Google Scholar · View at MathSciNet
  21. H. Zhao and L. Wang, “Hopf bifurcation in Cohen-Grossberg neural network with distributed delays,” Nonlinear Analysis: Real World Applications, vol. 8, no. 1, pp. 73–89, 2007. View at Publisher · View at Google Scholar · View at MathSciNet
  22. Q. Liu and W. Zheng, “Bifurcation of a Cohen-Grossberg neural network with discrete delays,” Abstract and Applied Analysis, vol. 2012, Article ID 909385, 11 pages, 2012. View at Publisher · View at Google Scholar · View at MathSciNet
  23. E. Kaslik and S. Balint, “Complex and chaotic dynamics in a discrete-time-delayed Hopfield neural network with ring architecture,” Neural Networks, vol. 22, no. 10, pp. 1411–1418, 2009. View at Publisher · View at Google Scholar · View at Scopus
  24. J. Hale, Theory of Functional Differential Equations, vol. 3 of Applied Mathematical Sciences, Springer, New York, NY, USA, 2nd edition, 1977. View at MathSciNet
  25. B. D. Hassard, N. D. Kazarinoff, and Y. H. Wan, Theory and Applications of Hopf bifurcation, vol. 41 of London Mathematical Society Lecture Note Series, Cambridge University Press, Cambridge, UK, 1981. View at MathSciNet
  26. R. S. Varga, Matrix Iterative Analysis, vol. 27 of Springer Series in Computational Mathematics, Springer, Berlin, Germany, 2000. View at Publisher · View at Google Scholar · View at MathSciNet