<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content="text/html; charset=iso-8859-1" http-equiv=Content-Type>
<META name=GENERATOR content="MSHTML 8.00.6001.18999"></HEAD>
<BODY>
<DIV><SPAN class=091431309-31012011><FONT face=Arial>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>Dear All,<?xml:namespace prefix = o ns =
"urn:schemas-microsoft-com:office:office" /><o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB> <o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>I 'm currently testing the optim macro in Scilab/FEA (external)
solver FEA<SPAN style="mso-spacerun: yes"> </SPAN>and I would like here to
share some feedbacks ...<o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB><o:p> </o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>First of all regarding the documentation, I noticed
:<o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB> <o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB><o:p> </o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB> <o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><B
style="mso-bidi-font-weight: normal"><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>A) <U>‘gc’ and bounds use</U><o:p></o:p></SPAN></B></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>'qn' -> BFGS algorithm<o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>'gc' -> conjugate gradient<o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>'nd' -> ?<o:p></o:p></SPAN></P>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>From the documentation <SPAN
style="mso-spacerun: yes"> </SPAN>"<B
style="mso-bidi-font-weight: normal">Note that the conjugate gradient solver
does not accept bounds on <CODE><SPAN
style="FONT-FAMILY: Arial">x</SPAN></CODE>. "</B><o:p></o:p></SPAN></P>
<P style="MARGIN: auto 0cm" class=para><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>Whereas the example "Scilab function" uses bounds with 'gc' ... ????
<o:p></o:p></SPAN></P>
<P style="MARGIN: auto 0cm" class=para><SPAN class=comment><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB>// Upper and lower bounds on x and setting up the algorithm to
'gc'</SPAN></SPAN><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB><BR><SPAN class=openclose>[</SPAN><SPAN class=id>f</SPAN><SPAN
class=default>,</SPAN><SPAN class=id>xopt</SPAN><SPAN
class=default>,</SPAN><SPAN class=id>gopt</SPAN><SPAN
class=openclose>]</SPAN><SPAN class=operator>=</SPAN><SPAN
class=command>optim</SPAN><SPAN class=openclose>(</SPAN><SPAN
class=functionid>cost</SPAN><SPAN class=default>,</SPAN><SPAN
class=string>'b'</SPAN><SPAN class=default>,</SPAN><SPAN
class=openclose>[</SPAN><SPAN class=operator>-</SPAN><SPAN
class=number>1</SPAN><SPAN class=default>;</SPAN><SPAN
class=number>0</SPAN><SPAN class=default>;</SPAN><SPAN
class=number>2</SPAN><SPAN class=openclose>]</SPAN><SPAN
class=default>,</SPAN><SPAN class=openclose>[</SPAN><SPAN
class=number>0.5</SPAN><SPAN class=default>;</SPAN><SPAN
class=number>1</SPAN><SPAN class=default>;</SPAN><SPAN
class=number>4</SPAN><SPAN class=openclose>]</SPAN><SPAN
class=default>,</SPAN><SPAN class=id>x0</SPAN><SPAN class=default>,</SPAN><SPAN
class=string>'gc'</SPAN><SPAN class=openclose>)</SPAN><o:p></o:p></SPAN></P><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2> <o:p></o:p></FONT></SPAN></PRE><PRE><FONT size=2><SPAN class=openclose><B style="mso-bidi-font-weight: normal"><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB>B) <U>bounds violated</U></SPAN></B></SPAN><B style="mso-bidi-font-weight: normal"><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p></o:p></SPAN></B></FONT></PRE><PRE><SPAN class=openclose><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>I tried several syntaxes for the bounds (with or without the dot coma i.e vector or its transposed) and all the time I noticed negative values of x …<o:p></o:p></FONT></SPAN></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><FONT size=2><SPAN style="FONT-FAMILY: Arial; COLOR: black; mso-ansi-language: EN-GB" lang=EN-GB>[fopt,optimized_parameters</SPAN><SPAN style="FONT-FAMILY: Arial; COLOR: #4a55db; mso-ansi-language: EN-GB" lang=EN-GB>]</SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB> <SPAN style="COLOR: #5c5c5c">=</SPAN> <SPAN style="COLOR: #32b9b9">optim</SPAN><SPAN style="COLOR: #4a55db">(</SPAN><SPAN style="COLOR: black">optimization_BFGS,</SPAN><SPAN style="COLOR: rosybrown">'b'</SPAN><SPAN style="COLOR: black">,</SPAN><SPAN style="COLOR: #4a55db">[</SPAN><SPAN style="COLOR: rosybrown">0</SPAN><SPAN style="COLOR: black">;</SPAN><SPAN style="COLOR: rosybrown">0</SPAN><SPAN style="COLOR: #4a55db">]</SPAN><SPAN style="COLOR: black">,</SPAN><SPAN style="COLOR: #4a55db">[</SPAN><SPAN style="COLOR: rosybrown">2</SPAN><SPAN style="COLOR: black">;</SPAN><SPAN style="COLOR: rosybrown">2</SPAN><SPAN style="COLOR: #4a55db">]</SPAN><SPAN style="COLOR: black">,initial_parameters</SPAN><SPAN style="COLOR: #5c5c5c">'</SPAN><SPAN style="COLOR: black">,</SPAN><SPAN style="COLOR: rosybrown">'ar'</SPAN><SPAN style="COLOR: black">,nap,iter,epsg,imp</SPAN><SPAN style="COLOR: #5c5c5c">=-</SPAN><SPAN style="COLOR: rosybrown">1</SPAN><SPAN style="COLOR: #4a55db">)</SPAN><SPAN style="COLOR: black">;<o:p></o:p></SPAN></SPAN></FONT></PRE><PRE><SPAN style="FONT-FAMILY: Arial; COLOR: black; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><B style="mso-bidi-font-weight: normal"><SPAN style="FONT-FAMILY: Arial; COLOR: black; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>Is there something wrong ?<o:p></o:p></FONT></SPAN></B></PRE><PRE><SPAN style="FONT-FAMILY: Arial; COLOR: black; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; COLOR: black; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><FONT size=2><SPAN class=openclose><B style="mso-bidi-font-weight: normal"><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB>B) <U>derivative use</U></SPAN></B></SPAN><B style="mso-bidi-font-weight: normal"><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p></o:p></SPAN></B></FONT></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2> </FONT></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>I noticed the strong influence of the h parameters (for the step) as well as the order parameter (order 2 by default) :<o:p></o:p></FONT></SPAN></PRE><PRE style="TEXT-INDENT: -18pt; MARGIN-LEFT: 71pt; mso-list: l0 level1 lfo1; tab-stops: 45.8pt list 71.0pt left 91.6pt 137.4pt 183.2pt 229.0pt 274.8pt 320.6pt 366.4pt 412.2pt 458.0pt 503.8pt 549.6pt 595.4pt 641.2pt 687.0pt 732.8pt"><SPAN style="FONT-FAMILY: Wingdings; mso-ansi-language: EN-GB; mso-fareast-font-family: Wingdings; mso-bidi-font-family: Wingdings" lang=EN-GB><SPAN style="mso-list: Ignore"><FONT size=2>ü</FONT><SPAN style="FONT: 7pt 'Times New Roman'"> </SPAN></SPAN></SPAN><FONT size=2><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB>If h is too small </SPAN><SPAN style="FONT-FAMILY: Wingdings; mso-ansi-language: EN-GB; mso-ascii-font-family: Arial; mso-hansi-font-family: Arial; mso-bidi-font-family: Arial; mso-char-type: symbol; mso-symbol-font-family: Wingdings" lang=EN-GB><SPAN style="mso-char-type: symbol; mso-symbol-font-family: Wingdings">è</SPAN></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB> a lot of evaluations<o:p></o:p></SPAN></FONT></PRE><PRE style="TEXT-INDENT: -18pt; MARGIN-LEFT: 71pt; mso-list: l0 level1 lfo1; tab-stops: 45.8pt list 71.0pt left 91.6pt 137.4pt 183.2pt 229.0pt 274.8pt 320.6pt 366.4pt 412.2pt 458.0pt 503.8pt 549.6pt 595.4pt 641.2pt 687.0pt 732.8pt"><SPAN style="FONT-FAMILY: Wingdings; mso-ansi-language: EN-GB; mso-fareast-font-family: Wingdings; mso-bidi-font-family: Wingdings" lang=EN-GB><SPAN style="mso-list: Ignore"><FONT size=2>ü</FONT><SPAN style="FONT: 7pt 'Times New Roman'"> </SPAN></SPAN></SPAN><FONT size=2><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB>If h is too high </SPAN><SPAN style="FONT-FAMILY: Wingdings; mso-ansi-language: EN-GB; mso-ascii-font-family: Arial; mso-hansi-font-family: Arial; mso-bidi-font-family: Arial; mso-char-type: symbol; mso-symbol-font-family: Wingdings" lang=EN-GB><SPAN style="mso-char-type: symbol; mso-symbol-font-family: Wingdings">è</SPAN></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB> loops quickly stop far from the optimized values <o:p></o:p></SPAN></FONT></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>order is essentially used for the function evaluation ; I tried different values of “order” thinking in a better function evaluation (for a better gradient calculation) … but the default value seems to be the best one.<o:p></o:p></FONT></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN></PRE><PRE><B style="mso-bidi-font-weight: normal"><U><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>C) Other algorithms<o:p></o:p></FONT></SPAN></U></B></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>I tested on a basic Rosenbrock function other order 0 method such as Genetic Algorithm or Simulated annealing but many function evaluations are necessary (more than a thousand) …. but probably the parameters have to be optimized themselves (Population size / number of mutations and so on ) … <o:p></o:p></FONT></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>since one evaluation need one FE modelling (i.e. the longest calculation duration), I’m not sure they are the most efficient for my application … finally the Nelder-Mead method is probably one of the best !<o:p></o:p></FONT></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>Regards<o:p></o:p></FONT></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><o:p><FONT size=2> </FONT></o:p></SPAN></PRE><PRE><SPAN style="FONT-FAMILY: Arial; mso-ansi-language: EN-GB" lang=EN-GB><FONT size=2>Paul<o:p></o:p></FONT></SPAN></PRE>
<P style="MARGIN: 0cm 0cm 0pt" class=MsoNormal><SPAN
style="FONT-FAMILY: Arial; FONT-SIZE: 10pt; mso-ansi-language: EN-GB"
lang=EN-GB><o:p> </o:p></SPAN></P></FONT></SPAN><SPAN
class=091431309-31012011><SPAN class=091431309-31012011><SPAN
class=openclose><FONT
face=Calibri></FONT></SPAN></DIV></SPAN></SPAN></BODY></HTML>
<pre>--------------------------------------------------------------------------------
Le présent mail et ses pièces jointes sont confidentiels et destinés à la personne ou aux personnes visée(s) ci-dessus. Si vous avez reçu cet e-mail par erreur, veuillez contacter immédiatement l'expéditeur et effacer le message de votre système. Toute divulgation, copie ou distribution de cet e-mail est strictement interdite.
This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error, please contact the sender and delete the email from your system. If you are not the named addressee you should not disseminate, distribute or copy this email.