In this paper, we present a method for approximating the solution of initial value ordinary differential equations with a priori error bounds. The method is based on a Chebyshev perturbation of the original differential equation together with the Frobenius method for solving the equation. Chebyshev
Constructive polynomial approximation with a priori error bounds for nonlinear initial value differential problems
✍ Scribed by L. Jódar; A.E. Posso; H. Castejón
- Publisher
- Elsevier Science
- Year
- 1999
- Tongue
- English
- Weight
- 486 KB
- Volume
- 38
- Category
- Article
- ISSN
- 0898-1221
No coin nor oath required. For personal study only.
✦ Synopsis
In this paper, we consider an initial value problem y' = fi x, y), yi 0) = Y0, where f is a continuous function satisfying a Lipschitz condition. First, the function fix, y) is approximated by a Bernstein polynomial in two variables, Bnif; x,y), of an appropriate degree according to a prescribed accuracy. Second, by application of the Frobenius method to the initial value problem y' = B,~(f; z, y), y(O) ----Y0, an exact series solution of the latter problem is constructed. Finally, the infinite series solution is truncated in order to obtain an explicit polynomial whose error with respect to the exact solution of the original problem is less than the prescribed accuracy, in an existence and uniqueness domain which is determined in terms of the initial data. (~) 1999 Elsevier Science Ltd. All rights reserved.
📜 SIMILAR VOLUMES
## In this paper, we construct polynomial approximate solutions with a priori error bounds of first-order quasi-linear initial-value partial differential problems with analytic Cauchy data. After constructing a power series solution, this is appropriately truncated according with a prefixed accura
The aim of this paper is double. First, we point out that the hypothesis D(tl)D(t2) = D(t2)D(tl) imposed in [1] can be removed. Second, a constructive method for obtaining analyticnumerical solutions with a prefixed accuracy in a bounded domain gl(to, tl) = [0,p] x [t0,tl], for mixed problems of the