Conjecture on prolate spheroidal quadrature error for exponential functions

About 13 years old · traced to

Let c>0c>0 and a≥0a\geq 0 be real numbers, and let n>2c/πn>2c/\pi be an integer. Let δn(eicax)\delta_n(e^{icax}) denote the quadrature error associated with the quadrature nodes tjt_j and weights WjW_j, and let λn\lambda_n be the prolate spheroidal eigenvalue. Exponential quadrature-error conjecture. If 0≤a≤10\leq a\leq 1, then

δn(eicax)=∣∫−11eicax dx−∑j=1neicatjWj∣≈∣λn∣2n.\delta_n\left(e^{icax}\right)=\left|\int_{-1}^1e^{icax}\,dx-\sum_{j=1}^ne^{icat_j}W_j\right|\approx |\lambda_n|^2\sqrt{n}.

If 1<a≤21<a\leq 2, then

δn(eicax)=∣∫−11eicax dx−∑j=1neicatjWj∣≈∣λn∣.\delta_n\left(e^{icax}\right)=\left|\int_{-1}^1e^{icax}\,dx-\sum_{j=1}^ne^{icat_j}W_j\right|\approx |\lambda_n|.

This conjecture summarizes numerical observations about the size of the quadrature error in the two ranges of the frequency parameter aa.

References

Primary source

Andrei Osipov and Vladimir Rokhlin, “On the evaluation of prolate spheroidal wave functions and associated quadrature rules”, arXiv:1301.1707 (2013).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.