Abstract:Summary.The trapezoidal rule with deferred corrections using uncentered end formulas is shown to converge. While the proof technique is more specialized than the standard asymptotic expansion approach, it has some advantages. In addition to providing a more complete theoretical justification for current implementations of deferred corrections with the trapezoidal rule, the approach given here will hopefully apply for several other discretization methods.
“…They suggest how this might be proved but do not follow through because "Such a proof would be quite tedious." In this paper we sketch a proof of this fact, which we believe is less tedious than that of Christiansen and Russell (1979) due to the way in which we break down the proof into smaller sinpZy stated results.…”
Section: Error Estimation and Iterative Improvement For The Numericalmentioning
confidence: 86%
“…(Lindberg; Van Rosendale, Skeel) 3. The Order of Accuracy for a Deferred Corrections Algorithm (Skeel) This work will constitute the future paper mentioned in Skeel (1980, section 4, second paragraph) in which we give an error analysis for a sequence of iterations for the algorithm considered by Christiansen and Russell (1979). This is intended to be a realistic example of the application of the theoretical framework for proving accuracy results, which because of its length was not included in .…”
Section: Multistep Methods For Ordinary Differential Equationsmentioning
confidence: 99%
“…satisfies the assumptions of Christiansen and Russell (1979) so that in particular the operator equation F(y) 0 has an isolated solution y.…”
Section: The Numerical Method Assume That the Differential Operator Fmentioning
confidence: 99%
“…In Christiansen and Russell (1979) a careful analysis of deferred corrections that does not involve asymptotic expansions is done for a realistic algorithm similar to the implementation of Lentini and Pereyra (1977) of iterated deferred corrections for two-point boundary value…”
Section: Error Estimation and Iterative Improvement For The Numericalmentioning
“…They suggest how this might be proved but do not follow through because "Such a proof would be quite tedious." In this paper we sketch a proof of this fact, which we believe is less tedious than that of Christiansen and Russell (1979) due to the way in which we break down the proof into smaller sinpZy stated results.…”
Section: Error Estimation and Iterative Improvement For The Numericalmentioning
confidence: 86%
“…(Lindberg; Van Rosendale, Skeel) 3. The Order of Accuracy for a Deferred Corrections Algorithm (Skeel) This work will constitute the future paper mentioned in Skeel (1980, section 4, second paragraph) in which we give an error analysis for a sequence of iterations for the algorithm considered by Christiansen and Russell (1979). This is intended to be a realistic example of the application of the theoretical framework for proving accuracy results, which because of its length was not included in .…”
Section: Multistep Methods For Ordinary Differential Equationsmentioning
confidence: 99%
“…satisfies the assumptions of Christiansen and Russell (1979) so that in particular the operator equation F(y) 0 has an isolated solution y.…”
Section: The Numerical Method Assume That the Differential Operator Fmentioning
confidence: 99%
“…In Christiansen and Russell (1979) a careful analysis of deferred corrections that does not involve asymptotic expansions is done for a realistic algorithm similar to the implementation of Lentini and Pereyra (1977) of iterated deferred corrections for two-point boundary value…”
Section: Error Estimation and Iterative Improvement For The Numericalmentioning
“…Of course the key theoretical problem is to show that can be used h y to approximate T 1 y to 0(h 2 ), and similarly for the higher-order T k , and On the other hand defect correction [3,9,12,13] …”
We consider the numerical solution of systems of nonlinear two point boundary value problems by Galerkin's method. An initial solution is computed with piecewise linear approximating functions and this is then improved by using higher-order piecewise polynomials to compute defect corrections. This technique, including numerical integration, is justified by typical Galerkin arguments and properties of piecewise polynomials rather than the traditional asymptotic error expansions of finite difference methods.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.