Abstract

Many mathematical models of complex processes may be posed as integro-differential equations (IDE). Many numerical methods have been proposed for solving those equations, but most of them are ad hoc thus new equations have to be solved from scratch for translating the IDE into the framework of the specific method chosen. Furthermore, there is a paucity of general-purpose numerical solvers that free the user from additional tasks. Here we present a general-purpose MATLAB (R) solver that has the above features. We have chosen to use a numerical quadrature algorithm combined with an accurate and efficient ODE solver both within a MATLAB environment to construct a routine (idsolver) capable of solving a wide variety of IDE of arbitrary order, including the Volterra and Fredholm IDE, variable limits on the integral, and non-linear IDE. The solver performs successive relaxation iterations until convergence is achieved. The user has to define a kernel, limits of integration and a forcing function, then launch the routine and get accurate results by tuning in a single tolerance parameter, as described below for several numerical examples. We have found, by solving several numerical examples from the literature, that the method is robust, fast and accurate. Program summary Program title: idsolver Catalogue identifier: AEQU_v1_0 Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEQU_v1_0.html Program obtainable from: CPC Program Library, Queen's University, Belfast, N. Ireland Licensing provisions: GNU General Public License No. of lines in distributed program, including test data, etc.: 372 No. of bytes in distributed program, including test data, etc.: 3435 Distribution format: tar.gz Programming language: MATLAB 2011b. Computer: PC, Macintosh. Operating system: Windows, OSX, Linux. RAM: 1 GB (1,073,741,824 bytes). Classification: 4.3, 4.11. Nature of problem: To solve a wide variety of integro-differential equations (IDE) of arbitrary order, including the Volterra and Fredholm IDE, variable limits on the integral, and non-linear IDE. Solution method: An efficient Lobatto quadrature, a robust and accurate IVP MATLAB's solver routine, and a recipe for combining old and new estimates that is equivalent to a successive relaxation method. Running time: The solver may take several seconds to execute. (C) 2013 Elsevier B.V. All rights reserved.