Abstract
To specify any such method, procedures need to be given for starting, for predicting and correcting, for stopping iterations on the corrector, and for rounding. Our purpose is to investigate corrector formulas. Briefly our idea is to study the propagated error associated with the many possible formulas, to single out certain families of formulas which should give small errors, and to compare individual members of these families in detail. All wellknown corrector formulas appear as special cases in the general setting of this approach. We restrict our attention to corrector formulas. In our experimental work we therefore iterate on these formulas until they are satisfied except for rounding errors. Our findings are therefore strictly applicable only to methods where such iterations are carried out, but our results may serve as a useful guide in other cases as well. We will write the corrector formula in the following general form,
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
