A note on the error growth in the numerical integration of periodic orbits

The aim of this note is to extend the analysis of B. Cano and J. M. Sanz-Serna [2] on the global error behaviour of general one step methods in the numerical integration of a periodic orbit to the case that such a periodic orbit can be embedded into a family of periodic orbits. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Families of periodic orbits in resonant reversible systems

We study the dynamics near an equilibrium point p ₀ of a Z ₂(ℝ)-reversible vector field in ℝ²ⁿ with reversing symmetry R satisfying R ² = I and dimFix(R) = n. We deal with one-parameter families of such systems X _λ such that X ₀ presents at p ₀ a degenerate resonance of type 0: p: q. We are assuming that the linearized system of X ₀ (at p ₀) has as eigenvalues: λ₁ = 0 and λ _j = ±iα _j , j = 2, … n. Our main concern is to find conditions for the existence of one-parameter families of periodic orbits near the equilibrium.     

Stable multistep methods of numerical integration of rigid systems of differential equations

We consider the basic propositions of the theory of multistep fractional-rational numerical methods with a variable step of integration. We establish general regularities in determining the coefficients of the methods. We prove the A-stability of these methods of arbitrary order, and also the absence of extraneous roots of the characteristic equations of the methods. Translated fromMatematichni Metodi ta Fiziko-Mekhanichni Polya, Vol. 39, No. 2, 1996, pp. 148–152.     

Error behaviour of multistep methods applied to unstable differential systems

The problem of modelling a dynamic system described by a system of ordinary differential equations which has unstable components for limited periods of time is discussed. It is shown that the global error in a multistep numerical method is the solution to a difference equation initial value problem, and the approximate solution is given for several popular multistep integration formulae. Inspection of the solution leads to the formulation of four criteria for integrators appropriate to unstable problems. A sample problem is solved numerically using three popular formulae and two different stepsizes to illustrate the appropriateness of the criteria.     

On the numerical integration of nonlinear initial value problems by linear multistep methods

We study the numerical solution of the nonlinear initial value problem $$\left\{ {\begin{array}{*{20}c} {{{du(t)} \mathord{\left/ {\vphantom {{du(t)} {dt}}} \right. \kern-\nulldelimiterspace} {dt}} + Au(t) = f(t),t > 0} \\ {u(0) = c,} \\ \end{array} } \right.$$ whereA is a nonlinear operator in a real Hilbert space. The problem is discretized using linear multistep methods, and we assume that their stability regions have nonempty interiors. We give sharp bounds for the global error by relating the stability region of the method to the monotonicity properties ofA. In particular we study the case whereAu is the gradient of a convex functional φ(u).     

A family of multistep methods to integrate orbits on spheres

Summary A trajectory problem is an initial value problem where the interest lies in obtaining the curve traced by the solution, rather than in finding the actual correspondence between the values of the parameter and the points on that curve. This paper introduces a family of multi-stage, multi-step numerical methods to integrate trajectory problems whose solution is on a spherical surface. It has been shown that this kind of algorithms has good numerical properties: consistency, stability, convergence and others that are not standard. The latest ones make them a better choice for certain problems.     

Blended extended linear multistep methods for the accurate numerical integration of stiff initial value problems

A class of blended extended linear multistep methods suitable for the approximate numerical integration of stiff systems of first order ordinary differential equations is described. These methods are formulated as a result of combining the second derivative extended backward differentiation formulae of Cash and the blended linear multistep methods of Skeel and Kong. The new methods combine a high order or accuracy with good stability properties and, as a direct consequence, they are often suitable for the numerical integration of stiff differential systems when high accuracy is requested. In the first part of the present paper we consider the derivation of these new blended methods and give the coefficients and stability regions for formulae of order up to and including 10. In the second half we consider their practical implementation. In particular we describe a variable order/variable step package based on these blended formulae and we evaluate the performance of this package on the well known DETEST test set. It is shown that the new code is reliable on this test set and is competitive with the well known second derivative method of Enright.     

A linear multistep numerical integration scheme for solving systems of ordinary differential equations with oscillatory solutions

In ^[1], a set of convergent and stable two-point formulae for obtaining the numerical solution of ordinary differential equations having oscillatory solutions was formulated. The derivation of these formulae was based on a non-polynomial interpolant which required the prior analytic evaluation of the higher order derivatives of the system before proceeding to the solution. In this paper, we present a linear multistep scheme of order four which circumvents this (often tedious) initial preparation. The necessary starting values for the integration scheme are generated by an adaptation of the variable order Gragg-Bulirsch-Stoer algorithm as formulated in ^[2].     

High order stiffly stable composite multistep methods for numerical integration of stiff differential equations

Let \(\dot y\) =f(y,t) withy(t ₀)=y ₀ possess a solutiony(t) fort≧t ₀. Sett _n=t ₀+nh, n=1, 2,.... Lety ₀ denote the approximate solution ofy(t _n) defined by the composite multistep method with \(\dot y_n \) =f(y _n ,t _n ) andN=1, 2,.... It is conjectured that the method is stiffly stable with orderp=l for alll≧1 and shown to be so forl=1,..., 25. The method is intrinsically efficient in thatl future approximate solution values are established simultaneously in an iterative solution process with only one function evaluation per iteration for each of thel future time points. Step and order control are easily implemented, in that the approximate solution at only one past point appears in each component multistep formula of the method and in that the local truncation error for the first component multistep formula of the method is easily evaluated as $$T^{[l]} = \frac{h}{{t_{Nl} - t_{(N - 1)l - 1} }}\{ y_{Nl}^{PRED} - y_{Nl} \} ,$$ wherey _Nl ^PRED denotes the value att _Nl of the Lagrange interpolating polynomial passing through the pointsy _(N?1)l+j att _(N?1)l+j withj=?1, 0,...,l ? 1.     

A stochastic approach to global error estimation in ODE multistep numerical integration

In order to assess the quality of approximate solutions obtained in the numerical integration of ordinary differential equations related to initial-value problems, there are available procedures which lead to deterministic estimates of global errors. The aim of this paper is to propose a stochastic approach to estimate the global errors, especially in the situations of integration which are often met in flight mechanics and control problems. Treating the global errors in terms of their orders of magnitude, the proposed procedure models the errors through the distribution of zero-mean random variables belonging to stochastic sequences, which take into account the influence of both local truncation and round-off errors. The dispersions of these random variables, in terms of their variances, are assumed to give an estimation of the errors. The error estimation procedure is developed for Adams-Bashforth-Moulton type of multistep methods. The computational effort in integrating the variational equations to propagate the error covariance matrix associated with error magnitudes and correlations is minimized by employing a low-order (first or second) Euler method. The diagonal variances of the covariance matrix, derived using the stochastic approach developed in this paper, are found to furnish reasonably precise measures of the orders of magnitude of accumulated global errors in short-term as well as long-term orbit propagations.     

High-order one-step P-stable methods for the numerical integration of periodic initial value problems

Using Lobatto nodes, one-step methods of order six and eight have been obtained for the second-order differential equation y″ = f(x, y), y(x₀) = y₀, y′(x₀) = y′₀. The methods are shown to be P-stable. If

, then at each integration step a system of dimension 3s, 4s, respectively, has to be solved. The numerical results, for two problems, obtained by using these methods are given in the end.     

Localization of periodic orbits of polynomial systems by ellipsoidal estimates

In this paper we study the localization problem of periodic orbits of multidimensional continuous-time systems in the global setting. Our results are based on the solution of the conditional extremum problem and using sign-definite quadratic and quartic forms. As examples, the Rikitake system and the Lamb’s equations for a three-mode operating cavity in a laser are considered.     

Homoclinic,heteroclinic and periodic orbits of singularly perturbed systems

The main aims of this paper are to study the persistence of homoclinic and heteroclinic orbits of the reduced systems on normally hyperbolic critical manifolds, and also the limit cycle bifurcations either from the homoclinic loop of the reduced systems or from a family of periodic orbits of the layer systems. For the persistence of homoclinic and heteroclinic orbits, and the limit cycles bifurcating from a homolinic loop of the reduced systems, we provide a new and readily detectable method to characterize them compared with the usual Melnikov method when the reduced system forms a generalized rotated vector field. To determine the limit cycles bifurcating from the families of periodic orbits of the layer systems, we apply the averaging methods.We also provide two four-dimensional singularly perturbed differential systems, which have either heteroclinic or homoclinic orbits located on the slow manifolds and also three limit cycles bifurcating from the periodic orbits of the layer system.     

Invariant manifolds and global error estimates of numerical integration schemes applied to stiff systems of singular perturbation type – Part II: Linear multistep methods

Summary. It is shown that appropriate linear multi-step methods (LMMs) applied to singularly perturbed systems of ODEs preserve the geometric properties of the underlying ODE. If the ODE admits an attractive invariant manifold so does the LMM. The continuous as well as the discrete dynamical system restricted to their invariant manifolds are no longer stiff and the dynamics of the full systems is essentially described by the dynamics of the systems reduced to the manifolds. These results may be used to transfer properties of the reduced system to the full system. As an example global error bounds of LMM-approximations to singularly perturbed ODEs are given. Received May 5, 1995 / Revised version received August 18, 1995     

Evaluation of the local error of fractional-rational multistep methods with variable integration step

To ensure the proper qualitative characteristic of approximate numerical solution of the Cauchy problem for a system of ordinary differential equations, it is necessary to formulate certain conditions that have to be satisfied by numerical methods. The efficiency of a numerical method is determined by constructing the algorithm of integration step changing and the choice of the order of the method. The construction of such a method requires one to determine preliminarily the admissible error of the method in each integration step. A theorem on the evaluation of the local error of multistep numerical p th-order methods with variable integration step without taking into account the round-off error is formulated. This theorem enables one to construct an efficient algorithm for the step change and the choice of the corresponding order of the method.     

An optimal control approach to the design of periodic orbits for mechanical systems with impacts

In this paper we study the problem of designing periodic orbits for a special class of hybrid systems, namely mechanical systems with underactuated continuous dynamics and impulse events. We approach the problem by means of optimal control. Specifically, we design an optimal control based strategy that combines trajectory optimization, dynamics embedding, optimal control relaxation and root finding techniques. The proposed strategy allows us to design, in a numerically stable manner, trajectories that optimize a desired cost and satisfy boundary state constraints consistent with a periodic orbit. To show the effectiveness of the proposed strategy, we perform numerical computations on a compass biped model with torso.