Analysis of a nonlinear singularly perturbed Volterra integro-differential equation
Introduction
Mathematical problems involving small parameter with the highest order derivative term are named as singularly perturbed problems. These problems are an important class of problems, because of their regular appearance in many applications in physics, biology, chemistry, and engineering. For small perturbation parameter, the solution of singularly perturbed problems exhibits multiscale character; within some parts of the domain the solution gradient is quite higher than in the other parts. This indicates that to get a convergent numerical solution on uniform meshes, we require the number of mesh points that is inversely proportional to the value of the perturbation parameter, which is computationally not feasible. Therefore, layer adaptive meshes are used to obtain accurate numerical solutions. Broadly there are two classes of such meshes. If a priori information about the location and width of the layers is known, a suitable layer adaptive mesh can be generated a priori (e.g. Shishkin mesh, Bakhvalov mesh, Vulanovic mesh, and so on [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]). But, many a times this is not the case. So, it is desirable to generate a suitable layer adaptive mesh a posteriori using some a posterior mesh generation algorithm [15], [16], [17], [18], [19], [20].
We consider following singularly perturbed nonlinear Volterra integro-differential equation (VIDE) where is a given constant and is a perturbation parameter which in general takes small values. The functions and are considered to be sufficiently smooth in and , respectively. Moreover, there exists an such that in . Such singularly perturbed VIDEs arise in various physical and biological systems, such as diffusion–dissipation processes, filament stretching problems, epidemic dynamics, and synchronous control systems [21], [22], [23], [24].
Numerical methods for a linear singularly perturbed VIDE are developed in [19], [25], [26], [27], [28], [29]. More specifically, an exponential type difference scheme is developed in [26]. In [28] the problem is solved using a fitted operator technique on a piecewise-uniform Shishkin mesh. In [27] a tension spline collocation method is presented. In [25] a backward difference formula is used for the derivative and a repeated quadrature rule is used for the integral term. Further, a Bakhvalov type mesh is used to resolve the layer. In [19] the integrand is considered to be with . Assuming the source term such that , a posteriori error estimate for a linear VIDE is derived. More precisely, it is proved that , where is the piecewise linear interpolant of the computed solution and is the backward difference operator. But surprisingly this a posteriori error estimate is not used for the adaptive mesh generation, instead an arc-length based monitor function is used. Numerical methods for a nonlinear singularly perturbed VIDE are developed in [30], [31]. In [30] the nonlinear problem with a special Kernel is solved by asymptotic expansions and an implicit Runge–Kutta method. In [31] a first order uniformly convergent finite difference scheme is constructed on a Bakhvalov type mesh.
To the best of our knowledge, no published paper derived a posteriori error estimate for nonlinear singularly perturbed VIDEs. Further, no published paper on (linear/nonlinear) singularly perturbed VIDEs derived a general criterion that can be used to immediately conclude parameter-uniform convergence of the scheme on a priori constructed layer adapted meshes. These gaps in the literature are the motivation of our work.
We discretize problem (1.1) by an implicit finite difference scheme on an arbitrary non-uniform mesh. The scheme comprises of an implicit difference operator for the derivative term and an appropriate quadrature rule for the integral term. We derive a priori error estimate in the discrete maximum norm based on which we can conclude parameter-uniform convergence of the scheme on a priori meshes, such as Shishkin and Bakhvalov meshes, in the same framework. More importantly, we derive a posteriori error estimate in the maximum norm, which can be used with any adaptive moving mesh procedure. We used a variant of de Boor’s algorithm [17], [18] for this purpose. Numerical experiments are performed and results are reported for validation of the theoretical error estimates.
The rest of the paper is structured as follows. The next section (Section 2), provides the stability result for the continuous problem (1.1). In Section 3, a finite difference discretization of problem (1.1) is described. A priori and a posteriori error estimates are derived in Sections 4 A priori error analysis, 5 A posteriori error analysis , respectively. In Section 6, we provide numerical results for validation of the theoretical error estimates. Finally, some conclusions are given in Section 7.
Notation: We use as a generic positive constant that does not depend upon and the discretization parameter. We simply denote the maximum norm for any function on the domain . The discrete maximum norm is denoted by , denotes the mesh on .
Section snippets
Stability of the continuous problem
This section provides the stability result for the continuous problem (1.1). It is used later in a posteriori error analysis of the present numerical scheme for problem (1.1).
Lemma 2.1 The solution of (1.1) satisfies the following stability estimate Further, if and are any two functions such that and where is a bounded piecewise continuous function. Then
Proof We rewrite problem (1.1) as follows
The discretization and its stability
We consider an arbitrary non-uniform mesh to discretize . We define and . Further, we define , for any mesh function . Now integrating (1.1) over , we get Using (1.1) and the right side rectangle formula we get where and We next approximate the integral term by the
A priori error analysis
The following bound on the derivative of the solution is needed for a priori error analysis, however it is not needed for a posteriori error analysis given in the next section.
Lemma 4.1 The solution of (1.1) satisfies [31]
Lemma 4.2 For the solutions and of (1.1) and (3.5), respectively, we have
Proof Using (3.3), (3.5), and the arguments in Lemma 3.2 the proof easily follows. □
Theorem 4.1 Suppose is the solution of (1.1) and is the solution of (3.5). Then
A posteriori error analysis
We now provide a posteriori error analysis for the discrete problem (3.5). Suppose is the piecewise linear interpolant of the numerical solution , so that is continuous on , linear on each , and . Further, for , we have
Theorem 5.1 Suppose is the solution of (1.1), is the solution of (3.5) on an arbitrary mesh , and is its piecewise linear interpolant. Then
Proof Using (1.1)
Numerical experiments
We consider the following test problem The exact solution of the test problem is given by . The quasilinearization technique for nonlinear problems is a Newton like method which gives an iterative scheme [31], [34] for the discrete problem (3.5) as follows where is given and We use the
Conclusions
A nonlinear singularly perturbed VIDE is considered. The discretization of the problem is done on an arbitrary non-uniform mesh by an implicit finite difference scheme which comprises of an implicit difference operator for the derivative term and an appropriate quadrature rule for the integral term. We have derived a priori error estimate in the discrete maximum norm based on which the scheme is proved to be uniformly convergent on Shishkin and Bakhvalov meshes in the same framework. After that
Acknowledgements
This research was supported by the Science and Engineering Research Board (SERB), India under the Project No.ECR/2017/000564. The first author gratefully acknowledges the support of University Grants Commission, India , for research fellowship with Reference No: 20/12/2015(ii)EU-V. The authors gratefully acknowledge the valuable comments and suggestions from the anonymous referees.
References (39)
- et al.
Numerical approximation of 2D time dependent singularly perturbed convection–diffusion problems with attractive or repulsive turning points
Appl. Math. Comput.
(2018) - et al.
Parameter uniform numerical method for singularly perturbed turning point problems exhibiting boundary layers
J. Comput. Appl. Math.
(2003) Layer-adapted methods for quasilinear singularly perturbed delay differential problems
Appl. Math. Comput.
(2014)- et al.
Parameter uniform optimal order numerical approximation of a class of singularly perturbed system of reaction diffusion problems involving a small perturbation parameter
J. Comput. Appl. Math.
(2019) Comparison of a priori and a posteriori meshes for singularly perturbed nonlinear parameterized problems
J. Comput. Appl. Math.
(2015)- et al.
Tension spline collocation methods for singularly perturbed Volterra integro-differential and Volterra integral equations
J. Comput. Appl. Math.
(2002) Implicit Runge-Kutta methods for some integrodifferential–algebraic equations
Appl. Numer. Math.
(1993)- et al.
Parameter-uniform approximation on equidistributed meshes for singularly perturbed parabolic reaction–diffusion problems with robin boundary conditions
Appl. Math. Comput.
(2021) On the optimization of the methods for solving boundary value problems in the presence of a boundary layer
Zh. Vychisl. Mat. Mat. Fiz.
(1969)Non-equidistant generalizations of the gushchin-shchennikov scheme
ZAMM-J. Appl. Math. Mech./Z. Angew. Math. Mech.
(1987)
A priori meshes for singularly perturbed quasilinear two-point boundary value problems
IMA J. Numer. Anal.
Robust Numerical Methods for Singularly Perturbed Differential Equations: Convection–Diffusion-Reaction and Flow Problems, Vol. 24
Fitted Numerical Methods for Singular Perturbation Problems
Numerical solution of nonlinear singularly perturbed problems on nonuniform meshes by using a non-standard algorithm
J. Math. Chem.
Numerical solution of time-fractional singularly perturbed convection–diffusion problems with a delay in time
Math. Methods Appl. Sci.
A second order uniformly convergent numerical scheme for parameterized singularly perturbed delay differential problems
Numer. Algorithms
Analysis of some numerical methods on layer adapted meshes for singularly perturbed quasilinear systems
Numer. Algorithms
Parameter-robust numerical method for a system of singularly perturbed initial value problems
Numer. Algorithms
Second order global uniformly convergent numerical method for a coupled system of singularly perturbed initial value problems
Appl. Math. Comput.
Cited by (9)
A priori and a posteriori error analysis for a system of singularly perturbed Volterra integro-differential equations
2023, Computational and Applied MathematicsA novel parameter-uniform numerical method for a singularly perturbed Volterra integro-differential equation
2023, Computational and Applied MathematicsA robust adaptive grid method for first-order nonlinear singularly perturbed Fredholm integro-differential equations
2023, Networks and Heterogeneous MediaError estimate of BDF2 scheme on a Bakhvalov-type mesh for a singularly perturbed Volterra integro-differential equation
2023, Networks and Heterogeneous MediaA Second-Order Adaptive Grid Method for a Singularly Perturbed Volterra Integrodifferential Equation
2022, Fractal and Fractional