The double-exponential methods need error estimates.
Closed this issue · 1 comments
emsr commented
They are recursive methods IIRC so it should be relatively easy.
emsr commented
Ok, so prev_sum has twice the step size as curr_sum. Also, if we hit max_iter we didn't keep te old prev_sum. So break if this is the last iter. Move n and step computation before the break. Do curr_sum - 2*prev_sum.