Optimal Order and Minimal Complexity of One-Step Methods for Initial Value Problems.
Abstract
We consider the task of numerically approximating the solution of an ordinary differential equation initial value problem. A methodology is given for determining the computational complexity of finding an approximate solution with error not exceeding epsilon. In addition, we determine the method of optimal order within a given class of methods, and show that under reasonable hypotheses, the optimal order increases as epsilon decreases, tending to infinity as epsilon tends to zero. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1976
- Accession Number
- ADA035931
Entities
People
- Arthur G. Werschulz
Organizations
- Carnegie Mellon University