Journal of negative results in Octave?

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Journal of negative results in Octave?

Jose Ramom Flores das Seixas

I would like to know if there is something like The negative results Journal for octave functions?

The reason I ask is because I implemented a quadrature for strongly oscillating functions, quadgF, but after comparing it with octave function quadgk I found, to my disappointment, that my function was not generally faster than quadgk.

The quadgk and quadl functions allowed to integrate strongly oscillating functions, but they are slow for my needs: perform diffraction calculations. Browsing the bibliography on the subject, I found an article by L.F. Shampine, reference [1], which describes an algorithm that seemed easy to implement to me. In fact, the proposed algorithm is an improvement of the SSP method [2,3], which possibly Mr. Shampine did not know, since he does not mention it in his bibliography.

In his article, Professor Shampine talks about a function he implemented, quadgF, based on the algorithm, but as much as I looked for it I couldn't find it -quadgF.m-. So I implemented it as well as I could, trying to vectorize it as much as possible, and kept the name used in [1].
It is quite possible that the reason why Professor Shampine's function has disappeared from the radar is precisely what I found after writing my implementation, i.e. that, in general, has no advantages over general functions like quadgk.
Anyway, for if my implementation served someone, either because he/she improved it and got more speed, or to avoid wasting time doing what I already did, I send attached the function I wrote, quadgF.m, and a document, quadgF.pdf, describing the algorithm and my implementation, as well as an analysis that compares my function and quadgk.
Yours sincerely
Jose Ramom




quadgF.m (8K) Download Attachment
quadgF.pdf (4M) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Journal of negative results in Octave?

Nir Krakauer-3
Looks like nice work. arXiv may be a good place to quickly post such a
result where anyone can find it, and a repository like BitBucket or
GitLab for making the program developed available.

On Mon, Sep 23, 2019 at 11:42 AM Jose Ramom Flores das Seixas
<[hidden email]> wrote:

>
> I would like to know if there is something like The negative results Journal for octave functions?
>
> The reason I ask is because I implemented a quadrature for strongly oscillating functions, quadgF, but after comparing it with octave function quadgk I found, to my disappointment, that my function was not generally faster than quadgk.
>
> The quadgk and quadl functions allowed to integrate strongly oscillating functions, but they are slow for my needs: perform diffraction calculations. Browsing the bibliography on the subject, I found an article by L.F. Shampine, reference [1], which describes an algorithm that seemed easy to implement to me. In fact, the proposed algorithm is an improvement of the SSP method [2,3], which possibly Mr. Shampine did not know, since he does not mention it in his bibliography.
>
> In his article, Professor Shampine talks about a function he implemented, quadgF, based on the algorithm, but as much as I looked for it I couldn't find it -quadgF.m-. So I implemented it as well as I could, trying to vectorize it as much as possible, and kept the name used in [1].
> It is quite possible that the reason why Professor Shampine's function has disappeared from the radar is precisely what I found after writing my implementation, i.e. that, in general, has no advantages over general functions like quadgk.
> Anyway, for if my implementation served someone, either because he/she improved it and got more speed, or to avoid wasting time doing what I already did, I send attached the function I wrote, quadgF.m, and a document, quadgF.pdf, describing the algorithm and my implementation, as well as an analysis that compares my function and quadgk.
> Yours sincerely
> Jose Ramom
>
> References
>
> 1Shampine L.F., Integrating oscillatory functions in MATLAB, II. Electronic Transactions on Numerical Analysis, volume 39, pp 403-413, 2012.
> 2Stamnes J.J., Spjelkavik B. & Pedersen H.M., Evaluation of diffraction integrals using local phase and amplitude approximations. Opt. Acta 30, 207-222, 1983.
> 3Stamnes J.J., Waves in focal regions: propagation, diffraction and focusing of light, sound and water waves. Adam Hilger, 1986.
>
>


Reply | Threaded
Open this post in threaded view
|

Re: Journal of negative results in Octave?

Jose Ramom Flores das Seixas

Às 18:03 de 23/09/19, Nir Krakauer escreveu:
> Looks like nice work. arXiv may be a good place to quickly post such a
> result where anyone can find it, and a repository like BitBucket or
> GitLab for making the program developed available.

Thank for your suggestions.




>
> On Mon, Sep 23, 2019 at 11:42 AM Jose Ramom Flores das Seixas
> <[hidden email]> wrote:
>> I would like to know if there is something like The negative results Journal for octave functions?
>>
>> The reason I ask is because I implemented a quadrature for strongly oscillating functions, quadgF, but after comparing it with octave function quadgk I found, to my disappointment, that my function was not generally faster than quadgk.
>>
>> The quadgk and quadl functions allowed to integrate strongly oscillating functions, but they are slow for my needs: perform diffraction calculations. Browsing the bibliography on the subject, I found an article by L.F. Shampine, reference [1], which describes an algorithm that seemed easy to implement to me. In fact, the proposed algorithm is an improvement of the SSP method [2,3], which possibly Mr. Shampine did not know, since he does not mention it in his bibliography.
>>
>> In his article, Professor Shampine talks about a function he implemented, quadgF, based on the algorithm, but as much as I looked for it I couldn't find it -quadgF.m-. So I implemented it as well as I could, trying to vectorize it as much as possible, and kept the name used in [1].
>> It is quite possible that the reason why Professor Shampine's function has disappeared from the radar is precisely what I found after writing my implementation, i.e. that, in general, has no advantages over general functions like quadgk.
>> Anyway, for if my implementation served someone, either because he/she improved it and got more speed, or to avoid wasting time doing what I already did, I send attached the function I wrote, quadgF.m, and a document, quadgF.pdf, describing the algorithm and my implementation, as well as an analysis that compares my function and quadgk.
>> Yours sincerely
>> Jose Ramom
>>
>> References
>>
>> 1Shampine L.F., Integrating oscillatory functions in MATLAB, II. Electronic Transactions on Numerical Analysis, volume 39, pp 403-413, 2012.
>> 2Stamnes J.J., Spjelkavik B. & Pedersen H.M., Evaluation of diffraction integrals using local phase and amplitude approximations. Opt. Acta 30, 207-222, 1983.
>> 3Stamnes J.J., Waves in focal regions: propagation, diffraction and focusing of light, sound and water waves. Adam Hilger, 1986.
>>
>>