Dynamics of a Nonstandard Finite-Difference Scheme for a Limit Cycle Oscillator with Delayed Feedback
Yuanyuan Wang1,2and Xiaohua Ding1
Academic Editor: Junjie Wei
Received06 May 2013
Accepted21 Jun 2013
Published24 Jul 2013
Abstract
We consider a complex autonomously driven single limit cycle oscillator with delayed feedback.
The original model is translated to a two-dimensional system. Through a nonstandard finite-difference (NSFD) scheme
we study the dynamics of this resulting system. The stability of the equilibrium of the model is investigated
by analyzing the characteristic equation. In the two-dimensional discrete model, we find that there are stability switches on the
time delay and Hopf bifurcation when the delay passes a sequence of critical
values. Finally, computer simulations are performed to illustrate the
theoretical results. And the results show that NSFD scheme is better than the Euler method.
1. Introduction
Reddy et al. [1] proposed the following model system of an autonomously driven single limit cycle oscillator:
where is a complex quantity, is the frequency of oscillation, and is a real constant. is the time delay of the autonomous feedback term. and represent the strengths of the linear and nonlinear contributions of the feedback. Reddy et al. investigated the temporal dynamics of system (1) in various regimes characterized by the natural parameters of the oscillator, strengths of the feedback components , and the time delay parameter . Jiang and Wei [2] have studied the stability of system (1) and have drawn the bifurcation diagram in plane in continuous-time model. Furthermore, it is found that there are stability switches on the time delay and Hopf bifurcation when the time delay crosses through some critical values.
But due to scientific computation and simulation, our interest focuses on the behavior of discrete dynamical system corresponding to (1). It is desired that the discrete-time model is “dynamically consistent” with the continuous-time model. In [3–12], the dynamics of numerical discrete difference equations can inherit those of the original differential equations. Wulf and Ford [13, 14] showed that the Euler forward method is “dynamically consistent” when applying it to solve the delay differential equation. It means that, for sufficiently small step-size, the discrete model undergoes a Hopf bifurcation of the same type with the original model.
In this paper, we apply NSFD scheme [15–17] to discretize (1). We consider the autonomous delay differential equation
The first-order derivative is approximated by modified forward Euler expression
with the denominator function such that
where stands for step-size and denotes the approximate value to , so we get the method as follows:
This method can seem as a modified forward Euler method [4]. NSFD scheme [15–17] tries to preserve the significant properties of their continuous analogues and consequently gives reliable numerical results. For small step-size we obtain the consistent dynamical results of the corresponding continuous-time model using Hopf bifurcation theory for discrete system [3, 6, 7, 13, 18]. Through the analysis, our results show that NSFD scheme is better than Euler method. To the best of our knowledge, to this day, by NSFD scheme, there are few results dealing with behavior of stability switches in the discrete model.
The paper is organized as follows. In Section 2, we discuss the distribution of the characteristic equation associated with the discrete limit cycle oscillator with delayed feedback and obtain the existence of the local Hopf bifurcation and stability switches. In Section 3, some computer simulations are performed to illustrate the theoretical results. In the final section, we summarize our results and give our future plans.
2. Stability Analysis
Let . Then (1) becomes
Set . Then (6) can be rewritten as
Let . Given , where , employ the NSFD scheme [15–17] to (7) and choose the “denominator function” as
It yields the difference equation
where
Introducing a new variable , we can rewrite (9) as
where , and
It is clear that the zero solution is a fixed point of (11), and the linearization of (11) around is
where
where is a unit matrix. The characteristic equation of is given by
Rewrite (15) in the more compact form
Similar to [19], we only need to investigate
Lemma 1. If , then all roots of (17) have modulus less than one for sufficiently small ; if , in all roots of (17), there is at least one root with modulus more than one.
Proof. For , (17) becomes
The equation has an -fold root and a simple root . Consider the root such that . This root is a function of . For (17), we have
Consequently, when , all roots of (17) lie in for sufficiently small ; when , in all roots of (17), there is at least one root with modulus more than one.
A Hopf bifurcation occurs when two roots of the characteristic equation (17) cross the unit circle. We have to find values of such that there exist roots on the unit circle. The roots on the unit circle are given by . Since we are dealing with a real polynomial complex roots, we only need to look for . is a root of (17) if and only if
Hence
Then
Or
So
If the step-size is sufficiently small, we obtain the following results.
Case I (k1> a). Consider the following:(I1) if , then , that is, ;(I2) if , then , that is, , which yields a contradiction;(I3), if , then , that is, , which yields a contradiction;(I4), if , then , that is, .
Case II (k1< a). Consider the following:(II1) if , then , that is, , which yields a contradiction;(II2) if , then , that is, ;(II3), if , then , that is, , which yields a contradiction;(II4), if , then , that is, .
Lemma 2. If the step-size is sufficiently small, and any one of (I2)(I3) and (II1)(II3) is satisfied, then (17) has no root with modulus one.
If the step-size is sufficiently small, any one of (I1)(I4) and (II2)(II4) (i.e., ) is satisfied, then . From (23) we know that
where .
It is clear that there exist values of the time delay parameters satisfying (23) according to .
Lemma 3. Let be a root of (17). If the step-size is sufficiently small and ((I1)(I4) and (II2)(II4)), then
Proof. From (17), we have
Substituting (25) into (28), we get
Therefore
If the step-size is sufficiently small and , we have
If the step-size is sufficiently small and , we have the following conclusions:
The proof is complete.
For the convenience, we denote (see the bifurcation diagram in [2])
Theorem 4.
For (11), the following statements are true.(1)If , then (11) undergoes a Hopf bifurcation at the origin when ; the zero solution is asymptotically stable when and is unstable when . (2)If , then (11) undergoes a Hopf bifurcation at the origin when . One controls (25) to make ( is determined in the proof) then the zero solution is asymptotically stable when and is unstable when .(3)If , then (11) undergoes a Hopf bifurcation at the origin when . We control (25) to guarantee the following statements.(3i) If , then the zero solution is unstable.(3ii) If , then the zero solution is asymptotically stable when and is unstable when .(4)If , then (11) undergoes a Hopf bifurcation at the origin when , and the zero solution is unstable.(5)If , then the zero solution of (11) is asymptotically stable.(6)If , then the zero solution of (11) is unstable.
Proof. (1) If , applying Lemmas 1 and 3, we know that all roots of (17) have modulus less than one when , and (17) has at least a couple of roots with modulus greater than one when . Due to Corollary 2.4 in Ruan and Wei [20], we get the conclusion. (2) If , the time delay satisfies
so
From (25) for , we know that . Since is a decreasing function for , we have . Meanwhile, in view of (34), we have
Set . If , then
There exists such that when , and when , . So we have when ; that is, . And when ; that is, ; this case is impossible. Applying Lemmas 1 and 3, we can arrive at the conclusion. (3) If , applying Lemmas 1 and 3, in the same way as (2), we can get the conclusion. (4) If , applying Lemmas 1 and 3, , we know that the zero solution is unstable. (5) (6) Applying Lemmas 1 and 2, we can get the conclusion.
According to the conclusions of Theorem 4, we have the results that are consistent with those for the corresponding continuous-time model, for sufficiently small step-size.
3. Numerical Simulations
One of the purposes of this section is to test the results in Section 2; the second one is to show that NSFD scheme is better than the Euler method.
We present some numerical results to system (7) with different values of , and . We choose ; the system (7) is given by
Using NSFD scheme () to (38), we obtain
Choosing , we obtaining Figures 2–5. We compute the bifurcation points of (39) for some step-size. We see that is asymptotically convergent to with the increasing of , which is the true value. Using NSFD scheme we have . Applying Euler method we obtain . From Figure 3 for , , we can obtain that the fixed point is not asymptotically stable. For and (Figure 4), the fixed point is asymptotically stable. With the Euler method (Figure 1) we can obtain that for , , the fixed point is not asymptotically stable. With NSFD scheme (Figure 2) for the same step-size and we can obtain that the fixed point is asymptotically stable.
(a)
(b)
(c)
(a)
(b)
(c)
(a)
(b)
(c)
(a)
(b)
(c)
(a)
(b)
(c)
Through the analysis, it demonstrates superiority of NSFD scheme over Euler method under the means of describing approximately the dynamics of the original system.
Analogous to the region , now we give the results and figures of in other five regions.
We choose . We have , and . See Figures 6, 7, and 8.
(a)
(b)
(c)
(a)
(b)
(c)
(a)
(b)
(c)
We choose . We have . A stability switch is found. See Figures 9, 10, and 11.
(a)
(b)
(c)
(a)
(b)
(c)
(a)
(b)
(c)
Choose . The zero solution of (39) is unstable. See Figure 12.
(a)
(b)
(c)
Choose . The zero solution of (39) is unstable. See Figure 13, where the delay is 1.
(a)
(b)
(c)
Take . The zero solution is asymptotically stable. See Figure 14.
(a)
(b)
(c)
Take . The zero solution is unstable. See Figure 15.
(a)
(b)
(c)
4. Conclusions and Future Plans
Reddy et al. [1] have studied the dynamics of a single Hopf bifurcation oscillator (the Stuart-Landau equation) in the presence of an autonomous time-delayed feedback. The feedback term has both a linear component and a simple quadratic nonlinear term. The model can also find more direct applications in simulation studies for feedback control of individual physical, chemical, or biological entities that have the basic nonlinear characteristics of Hopf oscillator.
In this article, we study a complex autonomously driven single limit cycle oscillator with delayed feedback. In Section 2, the original model is translated to a two-dimensional system. Using NSFD scheme we have investigated the dynamics of a discrete limit cycle oscillator with delayed feedback. Choose the “denominator function” as
Through analysis, we obtain Lemmas 1, 2, and 3 and Theorem 4. Equations (24) and (25) are important. For small step-size we obtain the consistent dynamical results of the corresponding continuous-time model. And we find stability switches in the two-dimensional discrete model. At the same time, it demonstrates superiority of NSFD scheme over the Euler method under the means of describing approximately the dynamics of the original system.
Consider a first-order complex differential equations with delay
where is a constant, and and are both complex. Cahlon and Schmidt [21] pointed out that (41) plays an important role as a test equation for studying the numerical method applied to delay differential equation. By studying the asymptotic stability of the solutions of (41) for different values of and , we learn more about the effect of the delay on the solution. For instance, if , the term with delay carries more weight, while the opposite holds for [21]. Wei and Zhang [19] have studied Hopf bifurcation and stability switches of (41) in the continuous-time model. In our future work, by many numerical methods, we will discuss those behaviors of (41) in the discrete-time model.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (11026189) and the NNSF of Shandong Province (no. ZR2010AQ021).
References
D. Reddy, A. Sen, and G. L. Johnston, “Dynamics of a limit cycle oscillator under time delayed linear and nonlinear feedbacks,” Physica D, vol. 144, no. 3-4, pp. 335–357, 2000.
W. Jiang and J. Wei, “Bifurcation analysis in a limit cycle oscillator with delayed feedback,” Chaos, Solitons and Fractals, vol. 23, no. 3, pp. 817–831, 2005.
X. Ding and H. Su, “Existence and convergence of Neimark-Sacker bifurcation for delay differential equations using Runge-Kutta methods,” International Journal of Computer Mathematics, vol. 88, no. 1, pp. 97–109, 2011.
H. Su and X. Ding, “Dynamics of a nonstandard finite-difference scheme for Mackey-Glass system,” Journal of Mathematical Analysis and Applications, vol. 344, no. 2, pp. 932–941, 2008.
Y. Wang, “Dynamics of a nonstandard finite-difference scheme for delay differential equations with unimodal feedback,” Communications in Nonlinear Science and Numerical Simulation, vol. 17, no. 10, pp. 3967–3978, 2012.
Q. Wang, D. Li, and M. Z. Liu, “Numerical Hopf bifurcation of Runge-Kutta methods for a class of delay differential equations,” Chaos, Solitons & Fractals, vol. 42, no. 5, pp. 3087–3099, 2009.
M. Liu and Q. Wang, “Numerical Hopf bifurcation of linear multistep methods for a class of delay differential equations,” Applied Mathematics and Computation, vol. 208, no. 2, pp. 462–474, 2009.
C. Zhang and B. Zheng, “Stability and bifurcation of a two-dimensional discrete neural network model with multi-delays,” Chaos, Solitons and Fractals, vol. 31, no. 5, pp. 1232–1242, 2007.
B. Zheng, Y. Zhang, and C. Zhang, “Stability and bifurcation of a discrete BAM neural network model with delays,” Chaos, Solitons and Fractals, vol. 36, no. 3, pp. 612–616, 2008.
C. Zhang, M. Liu, and B. Zheng, “Hopf bifurcation in numerical approximation of a class delay differential equations,” Applied Mathematics and Computation, vol. 146, no. 2-3, pp. 335–349, 2003.
X. Ding, H. Su, and M. Liu, “Stability and bifurcation of numerical discretization of a second-order delay differential equation with negative feedback,” Chaos, Solitons & Fractals, vol. 35, no. 4, pp. 795–807, 2008.
L. Zhang, C. Zhang, and D. Zhao, “Hopf bifurcation analysis of integro-differential equation with unbounded delay,” Applied Mathematics and Computation, vol. 217, no. 10, pp. 4972–4979, 2011.
V. Wulf and N. J. Ford, “Numerical Hopf bifurcation for a class of delay differential equations,” Journal of Computational and Applied Mathematics, vol. 115, no. 1-2, pp. 601–616, 2000.
V. Wulf, Numerical Analysis of Delay Differential Equations Undergoing A Hopf
Bifurcation, University of Liverpool, Liverpool, UK, 1999.
R. E. Mickens, “A nonstandard finite-difference scheme for the Lotka-Volterra system,” Applied Numerical Mathematics, vol. 45, no. 2-3, pp. 309–314, 2003.
K. C. Patidar, “On the use of nonstandard finite difference methods,” Journal of Difference Equations and Applications, vol. 11, no. 8, pp. 735–758, 2005.
L.-I. W. Roeger and R. E. Mickens, “Exact finite difference and non-standard finite difference schemes for ,” Journal of Difference Equations and Applications, vol. 18, no. 9, pp. 1511–1517, 2012.
J. Wei and C. Zhang, “Stability analysis in a first-order complex differential equations with delay,” Nonlinear Analysis. Theory, Methods & Applications. An International Multidisciplinary Journal. Series A: Theory and Methods, vol. 59, no. 5, pp. 657–671, 2004.
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 A, vol. 10, no. 6, pp. 863–874, 2003.
B. Cahlon and D. Schmidt, “On stability of a first-order complex delay differential equation,” Nonlinear Analysis. Real World Applications. An International Multidisciplinary Journal, vol. 3, no. 3, pp. 413–429, 2002.