Article Contents
Article Contents

# Parameterization method for unstable manifolds of delay differential equations

The second author was partially supported by NSF grant DMS - 1318172, NSF grant DMS- 1700154, and by the Alfred P. Sloan Foundation grant G-2016-7320.
• This work is concerned with efficient numerical methods for computing high order Taylor and Fourier-Taylor approximations of unstable manifolds attached to equilibrium and periodic solutions of delay differential equations. In our approach we first reformulate the delay differential equation as an ordinary differential equation on an appropriate Banach space. Then we extend the Parameterization Method for ordinary differential equations so that we can define operator equations whose solutions are charts or covering maps for the desired invariant manifolds of the delay system. Finally we develop formal series solutions of the operator equations. Order-by-order calculations lead to linear recurrence equations for the coefficients of the formal series solutions. These recurrence equations are solved numerically to any desired degree.

The method lends itself to a-posteriori error analysis, and recovers the dynamics on the manifold in addition to the embedding. Moreover, the manifold is not required to be a graph, hence the method is able to follow folds in the embedding. In order to demonstrate the utility of our approach we numerically implement the method for some 1, 2, 3 and 4 dimensional unstable manifolds in problems with constant, and (briefly) state dependent delays.

Mathematics Subject Classification: Primary: 34K28, 34K19; Secondary: 34C45, 37L25, 37M99, 37C27.

 Citation:

• Figure 1.  Parameterization of a 2D local unstable manifold attached to the origin for Wright's equation with $\alpha = 2.2$. The system has an attracting periodic orbit (yellow). We integrate an orbit (red) on the local unstable manifold until it converges to the periodic orbit. We compute the parameterization to Taylor order $K = 130$ by solving the recurrence equation (4). The coordinates used in the figure are discussed in Section A.1.3

Figure 2.  Boundary torus of a parameterized local unstable manifold attached to a periodic orbit for Wright's equation with $\alpha = 9$. We compute to Taylor order $K=42$ and Fourier order $M = 22$. The surface plotted in the frames is obtained by evaluating and plotting the image of the resulting Fourier-Taylor polynomial and exploits no numerical integration procedures. The torus is embedded in an infinite dimensional phase space so has no inside and outside. The coordinates used in the figure are discussed in Section A.1.

Figure 3.  The attractor (yellow) and the unstable manifold (red) of the cubic Ikeda equation at $u=0$. The attractor is computed simply by integrating arbitrary initial conditions using a standard DDE integrator. The unstable manifold is obtained by plotting the parameterization computed to 120 terms. Note that the parameterization captures several turns in the manifold, i.e. is far from the linear approximation (is not for example the graph of any function). The parameterization turns several times quite sharply near the ends

Figure 4.  Left frame: the first 9 eigenvalues of the linearized cubic Ikeda equation at the origin. Right frame: the decay of the coefficients of the parametrization of the unstable manifold at $u=0$

Figure 5.  Left frame: the first 8 eigenvalues of the linearized cubic Ikeda equation at the points $u=\pm1$. Right frame: the decay of the coefficients of the parametrization of the unstable manifolds at $u=\pm1$

Figure 6.  The equilibria (black dots), attractor (yellow) and two-dimensional manifolds at $u=\pm1$ (green) as seen from two different angles. Note how these two manifolds wrap closely around each other near the origin

Figure 7.  The decay of the coefficients of the parametrization of the unstable manifold of the minimal periodic solutions of the cubic Ikeda equation

Figure 8.  The attractor (yellow) and two-dimensional manifolds (blue) of the periodic orbits (magenta). The attractor is computed simply by integrating arbitrary initial conditions using a standard DDE integrator. The periodic orbit is computed using 140 Fourier modes. The unstable manifold is parameterized to Taylor order 51, where again each Taylor coefficient is computed with 140 Fourier modes

Figure 9.  Left frame: the first 8 eigenvalues of the linearized Wright's equation at the point $u=0$. Right frame: the decay of the coefficients of the parametrization of the unstable manifold at $u=0$

Figure 10.  The two-dimensional manifold at $u=0$ (green) as seen from two different angles

Figure 11.  Left frame: the first 10 eigenvalues of the linearized Wright's equation at the point $u=0$. Right frame: the decay of the coefficients of the parametrization of the unstable manifold at $u=0$

Figure 12.  The two-dimensional fast manifold (red) and the two-dimensional slow manifold (green) manifold at $u=0$ as seen from two different angles. The unstable periodic orbit of period $T \approx 0.805$ is shown in yellow. The manifolds seem to intersect one another (and the periodic orbit) because of the projection to 3 dimensions

Figure 13.  Several tori along the boundary of the 4D unstable manifold of the $u=0$ equilibrium of Wright's equation for $\alpha = 9.0$. Depicted are the tori corresponding to $c \in \{0.99, 0.61, 0.21, 0.08 \}$

Figure 14.  The decay of the coefficients computed for the parametrization of the unstable manifold, corresponding to the unstable periodic solution of Wright's equation

Figure 15.  The periodic orbit (yellow) and two-dimensional submanifold (blue) of the unstable manifold of the periodic orbit corresponding to $|z| = 0.25$

Figure 16.  Top frame: decay of the Taylor coefficients. Left frame: the scaling used for each $\epsilon$. Right frame: the defect in the conjugacy for each value of $\epsilon$

Figure 17.  The unstable manifold of the state-dependent perturbation of Wright's equation, for several values of $\epsilon$. The green manifold corresponds $\epsilon = 0$, while the dark blue one at the centre corresponds to $\epsilon = 1$. Note that while these manifolds do lie closely together, they do not exactly overlap

Figure 18.  Left frame: solution of a delay equation with initial history segment colored in red. Right frame: solution of same delay equation with the evolution of the initial segment by time $t = a$ colored red

Figure 19.  The phase space of a delay differential equation. The intuition is that we change to coordinates which "move with" the window of length $\tau$

Figure 20.  Left frame: the evolution of the delay embedding coordinates (green) corresponding to the solution of a delay equation with a given initial history (red). Right frame: the same delay embedding coordinates, but embedded in $\mathbb{R}^3$

Figure 21.  the dynamical meaning of the conjugacy described in equation (39). The desired chart map conjugates the dynamics on the unstable manifold to the linear flow generated by the unstable eigenvalues

Figure 22.  Unstable manifold of a periodic orbit $\gamma$

•  [1] A. Bellen and M. Zennaro, Numerical Methods for Delay Differential Equations, Numerical Mathematics and Scientific Computation, Oxford University Press, Oxford, 2013, First paperback reprint of the 2003 original [MR1997488]. [2] M. Breden, J. Lessard and J. Mireles James, Computation of maximal local (un)stable manifold patches by the parameterization method, Indagationes Mathematicae, 27 (2016), 340-367.  doi: 10.1016/j.indag.2015.11.001. [3] M. Breden, J.-P. Lessard and J. D. Mireles James, Computation of maximal local (un)stable manifold patches by the parameterization method, Indag. Math. (N.S.), 27 (2016), 340-367.  doi: 10.1016/j.indag.2015.11.001. [4] X. Cabré, E. Fontich and R. de la Llave, The parameterization method for invariant manifolds ⅰ: Manifolds associated to non-resonant subspaces, Indiana Univ. Math. J., 52 (2003), 283-328.  doi: 10.1512/iumj.2003.52.2245. [5] X. Cabré, E. Fontich and R. de la Llave, The parametrization method for invariant manifolds ⅱ: Regularity with respect to parameters, Indiana Univ. Math. J., 52 (2003), 283-328.  doi: 10.1512/iumj.2003.52.2245. [6] X. Cabré, E. Fontich and R. de la Llave, The parametrization method for invariant manifolds ⅲ: Overview and applications, Journal of Differential Equations, 218 (2005), 444-515. [7] R. C. Calleja, A. Celletti and R. de la Llave, A KAM theory for conformally symplectic systems: Efficient algorithms and their validation, J. Differential Equations, 255 (2013), 978-1049.  doi: 10.1016/j.jde.2013.05.001. [8] R. C. Calleja, T. Humphries and B. Krauskopf, Resonance phenomena in a scalar delay differential equation with two state-dependent delays Submitted, arXiv: 1607.02683. doi: 10.1137/16M1087655. [9] S. A. Campbell, Calculating center manifolds for delay differential equations using MapleTM, in Delay Differential Equations, Springer, New York, 2009,221-244. [10] M. J. Capiński, Computer assisted existence proofs of Lyapunov orbits at L2 and transversal intersections of invariant manifolds in the Jupiter-Sun PCR3BP, SIAM J. Appl. Dyn. Syst., 11 (2012), 1723-1753.  doi: 10.1137/110847366. [11] R. Castelli, J.-P. Lessard and J. Mireles-James, Parametrization of invariant manifolds for periodic orbits (ⅰ): Efficient numerics via the Floquet normal form, SIAM Journal on Applied Dynamical Systems, 14 (2015), 132-167.  doi: 10.1137/140960207. [12] R. Castelli, J. -P. Lessard and J. D. Mireles James, Parameterization of invariant manifolds for periodic orbits (ⅱ): a-posteriori analysis and computer assisted error bounds, To appear in Journal of Dynamics and Differential Equations. [13] G. M. Constantine and T. H. Savits, A multivariate Faà di Bruno formula with applications, Trans. Amer. Math. Soc., 348 (1996), 503-520.  doi: 10.1090/S0002-9947-96-01501-2. [14] R. de la Llave, A. González, À. Jorba and J. Villanueva, KAM theory without action-angle variables, Nonlinearity, 18 (2005), 855-895.  doi: 10.1088/0951-7715/18/2/020. [15] R. de la Llave and J. D. Mireles James, Connecting orbits for compact infinite dimensional maps: Computer assisted proofs of existence, SIAM Journal on Applied Dynamical Systems, 15 (2016), 1268-1323.  doi: 10.1137/15M1053608. [16] J. De Luca, N. Guglielmi, T. Humphries and A. Politi, Electromagnetic two-body problem: Recurrent dynamics in the presence of state-dependent delay, J. Phys. A, 43 (2010), 205103, 20. doi: 10.1088/1751-8113/43/20/205103. [17] M. Dellnitz, M. Hessel-Von Molo and A. Ziessler, On the computation of attractors for delay differential equations Submitted. doi: 10.3934/jcd.2016005. [18] K. Engelborghs, T. Luzyanina and D. Roose, Numerical bifurcation analysis of delay differential equations using DDE-BIFTOOL, ACM Trans. Math. Software, 28 (2002), 1-21.  doi: 10.1145/513001.513002. [19] G. Farkas, Unstable manifolds for RFDEs under discretization: the Euler method, Comput. Math. Appl., 42 (2001), 1069-1081, Numerical Methods and Computational Mechanics (Miskolc, 1998).  doi: 10.1016/S0898-1221(01)00222-X. [20] G. Farkas, Discretizing hyperbolic periodic orbits of delay differential equations, ZAMM Z. Angew. Math. Mech., 83 (2003), 38-49.  doi: 10.1002/zamm.200310003. [21] J. -L. Figueras, M. Gameiro, J. -P. Lessard and R. de la Llave, A framework for the numerical computation and a posteriori verification of invariant objects of evolution equations Submitted. doi: 10.1137/16M1073777. [22] H. Froehling, J. P. Crutchfield, D. Farmer, N. H. Packard and R. Shaw, On determining the dimension of chaotic flows, Phys. D, 3 (1981), 605-617.  doi: 10.1016/0167-2789(81)90043-9. [23] R. H. Goodman and J. K. Wróbel, High-order bisection method for computing invariant manifolds of two-dimensional maps, Internat. J. Bifur. Chaos Appl. Sci. Engrg., 21 (2011), 2017-2042.  doi: 10.1142/S0218127411029604. [24] K. Green, B. Krauskopf and K. Engelborghs, One-dimensional unstable eigenfunction and manifold computations in delay differential equations, J. Comput. Phys., 197 (2004), 86-98.  doi: 10.1016/j.jcp.2003.11.018. [25] A. Guillamon and G. Huguet, A computational and geometric approach to phase resetting curves and surfaces, SIAM J. Appl. Dyn. Syst., 8 (2009), 1005-1042.  doi: 10.1137/080737666. [26] J. Hale, Theory of Functional Differential Equations, 2nd edition, Applied Mathematical Sciences, Vol. 3, Springer-Verlag, New York-Heidelberg, 1977. [27] J. K. Hale and S. M. Verduyn Lunel, Introduction to Functional-Differential Equations, Applied Mathematical Sciences, 99, Springer-Verlag, New York, 1993. doi: 10.1007/978-1-4612-4342-7. [28] A. Haro, Automatic differentiation methods in computational dynamical systems: Invariant manifolds and normal forms of vector fields at fixed points, IMA Note. [29] A. Haro, M. Canadell, J. -L. Figueras, A. Luque and J. -M. Mondelo, The Parameterization Method for Invariant Manifolds. From Rigorous Results to Effective Computations, Applied Mathematical Sciences, 195, Springer, 2016. doi: 10.1007/978-3-319-29662-3. [30] À. Haro and R. de la Llave, A parameterization method for the computation of invariant tori and their whiskers in quasi-periodic maps: Numerical algorithms, Discrete Contin. Dyn. Syst. Ser. B, 6 (2006), 1261-1300 (electronic), doi: 10.3934/dcdsb.2006.6.1261. [31] A. Haro and R. de la Llave, A parameterization method for the computation of invariant tori and their whiskers in quasi-periodic maps: Rigorous results, J. Differential Equations, 228(2006), 530-579, doi: 10.1016/j.jde.2005.10.005. [32] A. Haro and R. de la Llave, A parameterization method for the computation of invariant tori and their whiskers in quasi-periodic maps: Explorations and mechanisms for the breakdown of hyperbolicity SIAM J. Appl. Dyn. Syst., 6(2007), 142-207 (electronic), doi: 10.1137/050637327. [33] F. Hartung, T. Krisztin, H. -O. Walther and J. Wu, Functional differential equations with state-dependent delays: Theory and applications, in Handbook of Differential Equations: Ordinary Differential Equations. Vol. III, Handb. Differ. Equ., Elsevier/North-Holland, Amsterdam, 2006,435-545. doi: 10.1016/S1874-5725(06)80009-X. [34] X. He and R. de la Llave, Construction of quasi-periodic solutions of state-dependent delay differential equations by the parameterization method Ⅰ: Finitely differentiable, hyperbolic case, hyperbolic case, Journal of Dynamics and Differential Equations (2016). Available from: https://www.ma.utexas.edu/mp_arc/c/15/15-105.pdf. [35] X. He and R. de la Llave, Construction of quasi-periodic solutions of state-dependent delay differential equations by the parameterization method Ⅱ: Analytic case, Journal of Differential Equations, 261 (2016), 2068-2108.  doi: 10.1016/j.jde.2016.04.024. [36] T. Heil, I. Fischer, W. Elsäßer, B. Krauskopf, A. Gavrielides and K. Green, Delay dynamics of semiconductor lasers with short external cavities: Bifurcation scenarios and mechanisms, Phys. Rev. E (3), 67(2003), 066214, 11. doi: 10.1103/PhysRevE.67.066214. [37] G. Huguet and R. de la Llave, Computation of limit cycles and their isochrons: Fast algorithms and their convergence, SIAM J. Appl. Dyn. Syst., 12 (2013), 1763-1802.  doi: 10.1137/120901210. [38] A. R. Humphries, O. A. DeMasi, F. M. G. Magpantay and F. Upham, Dynamics of a delay differential equation with multiple state-dependent delays, Discrete Contin. Dyn. Syst., 32 (2012), 2701-2727.  doi: 10.3934/dcds.2012.32.2701. [39] T. Johnson and W. Tucker, A note on the convergence of parametrised non-resonant invariant manifolds, Qual. Theory Dyn. Syst., 10 (2011), 107-121.  doi: 10.1007/s12346-011-0040-2. [40] G. S. Jones, The existence of periodic solutions of $f^{\prime} (x)=-α f(x-1)\{1+f(x)\}$, J. Math. Anal. Appl., 5 (1962), 435-450.  doi: 10.1016/0022-247X(62)90017-3. [41] G. S. Jones, On the nonlinear differential-difference equation $f^{\prime} (x)=-α f(x-1)\{1+f(x)\}$, J. Math. Anal. Appl., 4 (1962), 440-469.  doi: 10.1016/0022-247X(62)90041-0. [42] À. Jorba and M. Zou, A software package for the numerical integration of ODEs by means of high-order Taylor methods, Experiment. Math., 14 (2005), 99-117.  doi: 10.1080/10586458.2005.10128904. [43] G. Kiss and J.-P. Lessard, Computational fixed-point theory for differential delay equations with multiple time lags, J. Differential Equations, 252 (2012), 3093-3115.  doi: 10.1016/j.jde.2011.11.020. [44] G. Kiss and J. -P. Lessard, Rapidly and slowly oscillating periodic oscillations of a delayed van der pol oscillator, Submitted. [45] D. E. Knuth, The Art of Computer Programming. Vol. 2 Seminumerical Algorithms, Third edition [of MR0286318], Addison-Wesley, Reading, MA, 1998. [46] B. Krauskopf, H. Osinga, E. Doedel, M. Henderson, J. Guckenheimer, A. Vladimirsky, M. Dellnitz and O. Junge, A survey of methods for computing (un)stable manifolds of vector fields, Int. J. Bifurcation & Chaos, 15 (2005), 763-791.  doi: 10.1142/S0218127405012533. [47] B. Krauskopf and K. Green, Computing unstable manifolds of periodic orbits in delay differential equations, J. Comput. Phys., 186 (2003), 230-249.  doi: 10.1016/S0021-9991(03)00050-0. [48] J.-P. Lessard, Recent advances about the uniqueness of the slowly oscillating periodic solutions of Wright's equation, J. Differential Equations, 248 (2010), 992-1016.  doi: 10.1016/j.jde.2009.11.008. [49] J. -P. Lessard, Delay Differential Equations and Continuation, Lecture notes for the AMS short course on rigorous numerics in dynamics. [50] J.-P. Lessard, J. D. Mireles James and C. Reinhardt, Computer assisted proof of transverse saddle-to-saddle connecting orbits for first order vector fields, J. Dynam. Differential Equations, 26 (2014), 267-313.  doi: 10.1007/s10884-014-9367-0. [51] J. Lessard, J. D. Mireles James and J. Ransford, Automatic differentiation for fourier series and the radii polynomial approach To appear in Physica D. doi: 10.1016/j.physd.2016.02.007. [52] X. Li and R. de la Llave, Construction of quasi-periodic solutions of delay differential equations via kam techniques, Journal of Differential Equations, 247 (2009), 822-865.  doi: 10.1016/j.jde.2009.03.009. [53] J. Mireles-James and K. Mischaikow, Rigorous a-posteriori computation of (un)stable manifolds and connecting orbits for analytic maps, SIAM Journal on Applied Dynamical Systems, 12 (2013), 957-1006.  doi: 10.1137/12088224X. [54] J. D. Mireles James, Computer assisted error bounds for linear approximation of (un)stable manifolds and rigorous validation of higher dimensional transverse connecting orbits, Commun. Nonlinear Sci. Numer. Simul., 22 (2015), 1102-1133.  doi: 10.1016/j.cnsns.2014.08.010. [55] J. D. Mireles James, Fourier-taylor approximation of unstable manifolds for compact maps: Numerical implementation and computer assisted error bounds, Foundations of Computational Mathematics, (2016). doi: 10.1007/s10208-016-9325-9. [56] J. D. Mireles James and H. Lomelí, Computation of heteroclinic arcs with application to the volume preserving Hénon family, SIAM J. Appl. Dyn. Syst., 9 (2010), 919-953.  doi: 10.1137/090776329. [57] A. Neumaier and T. Rage, Rigorous chaos verification in discrete dynamical systems, Phys. D, 67 (1993), 327-346.  doi: 10.1016/0167-2789(93)90169-2. [58] R. D. Nussbaum, Periodic solutions of analytic functional differential equations are analytic, Michigan Math. J., 20 (1973), 249-255.  doi: 10.1307/mmj/1029001104. [59] N. H. Packard, J. P. Crutchfield, D. Farmer and R. Shaw, Geometry from a time series, Physical Review Letters, 45 (1980), 712-716.  doi: 10.1007/BF02650178. [60] C. Reinhardt and J. M. James, Fourier-taylor parameterization of unstable manifolds for parabolic partial differential equations: Formalism, implementation, and rigorous validation, Submitted, arXiv: 1601.00307. [61] C. Reinhardt, J. B. Van den Berg and J. D. Mireles James, Computing (un)stable manifolds with validated error bounds: Non-resonant and resonant spectra, Journal of Nonlinear Science, 26 (2016), 1055-1095.  doi: 10.1007/s00332-016-9298-5. [62] D. Roose, T. Luzyanina, K. Engelborghs and W. Michiels, Software for stability and bifurcation analysis of delay differential equations and applications to stabilization, in Advances in Time-Delay Systems, Lect. Notes Comput. Sci. Eng., 38, Springer, Berlin, 2004,167-181. doi: 10.1007/978-3-642-18482-6_12. [63] T. Sahai and A. Vladimirsky, Numerical methods for approximating invariant manifolds of delayed systems, SIAM J. Appl. Dyn. Syst., 8 (2009), 1116-1135.  doi: 10.1137/080718772. [64] G. Samaey, K. Engelborghs and D. Roose, Numerical computation of connecting orbits in delay differential equations, Numer. Algorithms, 30 (2002), 335-352.  doi: 10.1023/A:1020102317544. [65] T. Sauer, J. A. Yorke and M. Casdagli, Embedology, J. Statist. Phys., 65 (1991), 579-616.  doi: 10.1007/BF01053745. [66] J. Sieber and B. Krauskopf, Bifurcation analysis of an inverted pendulum with delayed feedback control near a triple-zero eigenvalue singularity, Nonlinearity, 17 (2004), 85-103.  doi: 10.1088/0951-7715/17/1/006. [67] J. C. Sprott, A simple chaotic delay differential equation, Phys. Lett. A, 366 (2007), 397-402.  doi: 10.1016/j.physleta.2007.01.083. [68] R. Szczelina and P. Zgliczyński, Algorithm for rigorous integration of delay differential equations and the computer-assisted proof of periodic orbits in the Mackey-Glass equation, Submitted, http://ww2.ii.uj.edu.pl/~zgliczyn/papers/delay/main-dde.pdf. [69] F. Takens, Detecting strange attractors in turbulence, in Dynamical Systems and Turbulence, Warwick 1980 (Coventry, 1979/1980), Lecture Notes in Math., 898, Springer, Berlin-New York, 1981,366-381. [70] J. B. van den Berg and J. D. Mireles James, Parametrization of slow-stable manifolds and their invariant vector bundles: Theory and numerical implementation, Discrete Contin. Dyn. Syst., 36 (2016), 4637-4664.  doi: 10.3934/dcds.2016002. [71] H.-O. Walther, Topics in delay differential equations, Jahresber. Dtsch. Math.-Ver., 116 (2014), 87-114.  doi: 10.1365/s13291-014-0086-6. [72] D. Wilczak and P. Zgliczyński, Heteroclinic connections between periodic orbits in planar restricted circular three body problem. Ⅱ, Comm. Math. Phys., 259 (2005), 561-576.  doi: 10.1007/s00220-005-1374-x. [73] E. M. Wright, A non-linear difference-differential equation, J. Reine Angew. Math., 194 (1955), 66-87.  doi: 10.1515/crll.1955.194.66. [74] J. K. Wróbel and R. H. Goodman, High-order adaptive method for computing two-dimensional invariant manifolds of three-dimensional maps, Commun. Nonlinear Sci. Numer. Simul., 18 (2013), 1734-1745.  doi: 10.1016/j.cnsns.2012.10.017.

Figures(22)