<html><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /></head><body style='font-size: 10pt; font-family: Verdana,Geneva,sans-serif'>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace">Of course I did :-)</div>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace"> </div>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace">It is a way as well to share feedback on the influence of some flags; as it has been said 'it is my understanding" ...</div>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace"> </div>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace">Paul</div>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace"> </div>
<div class="pre" style="margin: 0; padding: 0; font-family: monospace">Le 2017-01-23 11:37, Paul Bignier a écrit :
<blockquote type="cite" style="padding: 0 0.4em; border-left: #1010ff 2px solid; margin: 0">Hi Paul,<br /> <br /> As I'm sure you've read them, optim [<a href="https://help.scilab.org/docs/6.0.0/en_US/optim.html" target="_blank" rel="noreferrer">1</a>], neldermead [<a href="https://help.scilab.org/docs/6.0.0/en_US/neldermead.html" target="_blank" rel="noreferrer">2</a>] &<br /> numderivative [<a href="https://help.scilab.org/docs/6.0.0/en_US/numderivative.html" target="_blank" rel="noreferrer">3</a>] help pages provide you with all the flags you can<br /> customize.<br /> After that, it's all in the functions that you provide to compute the<br /> cost & its derivative.<br /> <br /> Best regards,<br /> Paul<br /> <br /> On 01/23/2017 09:41 AM, <a href="mailto:paul.carrico@free.fr">paul.carrico@free.fr</a> wrote:<br /> <br />
<blockquote type="cite" style="padding: 0 0.4em; border-left: #1010ff 2px solid; margin: 0">Hi All<br /> <br /> I’m using ‘optim’ and ‘NelderMead’ in conjunction with my<br /> finite element solver.<br /> <br /> A “good” optimization is a balance between accuracy and cpu time<br /> … in other word I do not necessary need to be accurate at 1e-11<br /> but requiring a lot of iterations where 1e-3 is enough with a lower<br /> amount of loops.<br /> <br /> In my understanding:<br /> <br /> * With ‘optim’, I can modifiy<br /> <br /> * The step value h in numderivative (put at 1e-3 after previous<br /> tests on analytical functions tests)<br /> * The values of epsf (default value) and epsg (tested at 1e-3 and<br /> 1e-5)<br /> <br /> * With “Neldermead”, I can change:<br /> <br /> * Tolfunrelative (tested at default value for the moment)<br /> * Tolxrelative (tested at default value for the moment)<br /> <br /> Am I right or is there another 'flags' I can play with?<br /> <br /> _NB_: so far, Nelder-Mead requires less iterations than ‘optim’<br /> with a single variable … I’m wondering how can I improve optim<br /> use that is supposed to converge faster?<br /> <br /> Thanks<br /> <br /> Paul<br /> <br /> _______________________________________________<br /> users mailing list<br /> <a href="mailto:users@lists.scilab.org">users@lists.scilab.org</a><br /> <a href="http://lists.scilab.org/mailman/listinfo/users" target="_blank" rel="noreferrer">http://lists.scilab.org/mailman/listinfo/users</a></blockquote>
<br /> -- <br /> Paul BIGNIER<br /> Development engineer<br /> -----------------------------------------------------------<br /> Scilab Enterprises<br /> 143bis rue Yves Le Coz - 78000 Versailles, France<br /> Phone: +33.1.80.77.04.68<br /> <a href="http://www.scilab-enterprises.com" target="_blank" rel="noreferrer">http://www.scilab-enterprises.com</a><br /> <br /> <br /> Links:<br /> ------<br /> [1] <a href="https://help.scilab.org/docs/6.0.0/en_US/optim.html" target="_blank" rel="noreferrer">https://help.scilab.org/docs/6.0.0/en_US/optim.html</a><br /> [2] <a href="https://help.scilab.org/docs/6.0.0/en_US/neldermead.html" target="_blank" rel="noreferrer">https://help.scilab.org/docs/6.0.0/en_US/neldermead.html</a><br /> [3] <a href="https://help.scilab.org/docs/6.0.0/en_US/numderivative.html" target="_blank" rel="noreferrer">https://help.scilab.org/docs/6.0.0/en_US/numderivative.html</a><br /> _______________________________________________<br /> users mailing list<br /> <a href="mailto:users@lists.scilab.org">users@lists.scilab.org</a><br /> <a href="http://lists.scilab.org/mailman/listinfo/users" target="_blank" rel="noreferrer">http://lists.scilab.org/mailman/listinfo/users</a></blockquote>
</div>
</body></html>