[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [amibroker] optimization



PureBytes Links

Trading Reference Links










Hi Ed and Dave –

 

I don’t think the simplex method is
applicable in this case.  Its use is to solve linear programming problems
by moving along the edge of the convex space defined by the constraining
inequalities.  If we had a well formed set of inequalities, and if we knew
the space they defined was convex, and if we knew the optimal solution was at
one of the vertices, then we might try to simplify the parameter search using the
simplex method or something similar.  Unfortunately, none of those three ifs
are satisfied for our model development problem.

 

I am in the process of designing and
programming a genetic search routine (as a dll plugin) that will look
throughout the search space by itself and keep track of the best solutions
found.  It avoids the problem of setting up a x by y by z optimization that
steps through all the possibilities in order.  I plan to be able to
specify any optimization complexity, start the processing using any reasonable values
for the optimization variables, and let the search run by itself.  When it
times out, counts out, or is interrupted, a list of the best solutions is
available for further testing.

 

When I get it working, and if it looks
promising, I’ll share the code.  But don’t hold your breath.

 

Howard

 

 

 



-----Original Message-----
From: Dave Merrill
[mailto:dmerrill@xxxxxxx] 
Sent: Friday, October 24, 2003
1:29 PM
To: amibroker@xxxxxxxxxxxxxxx
Subject: RE: [amibroker]
optimization

 



<span
>hadn't heard of
it before, but found some articles describing it. sounds like you're more
familiar than I am with this kind of thing, but if it's helpful, try:





 





<span
><a
href="">http://flash.lakeheadu.ca/~kliu/simplex.htm





 





<span
>clearly you're
right that reducing the number of combinations tried would improve processing
time. I'm particularly interested in this because my auto-optimization
framework needs to do this independently for each ticker.





 





<span
>I don't have the
time to code this reference up in AFL, and probably not the math expertise
either. if you do it, please let me know, and I'll look into adding it to the
auto-optimize thing.





 





<span
>dave





 





Since I am playing with the Optimization function right now I was <font
size=2 face="Courier New">
wondering the following. If you optimize a set op
parameters 
simultaneously than the Optimization procedure
calculates all 
possible combinations of variables. There are a
number of techniques 
that avoid this e.g. the simplex method. 

Did anyone work on such a function?

regards, Ed







<font size=2
face="Courier New">Send BUG REPORTS to
bugs@xxxxxxxxxxxxx<span
>
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to:
amiquote@xxxxxxxxxxxxxxx 
(Web page: <a
href="">http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: <a
href="">http://groups.yahoo.com/group/amibroker/files/groupfaq.html


Your use of
Yahoo! Groups is subject to the Yahoo!
Terms of Service. 












Yahoo! Groups Sponsor


  ADVERTISEMENT 









Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx 
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.