[SciPy-user] adaptive simulated annealing (ASA)
![](https://secure.gravatar.com/avatar/cf4af8309b03ebfc0e11ac65245f3efb.jpg?s=120&d=mm&r=g)
Hi there, The message 'adaptive simulated annealing (ASA)' is over four years old now, asking for a python implementation of ASA (http://www.ingber.com/#ASA). I was wondering if such an implementation is around today? Kind regards, Dieter -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
![](https://secure.gravatar.com/avatar/cf4af8309b03ebfc0e11ac65245f3efb.jpg?s=120&d=mm&r=g)
Dear reader, I give it another try with more explanation. If it remains unanswered I will assume that such an implementation does not exist. Adaptive Simulated Annealing (ASA) is a freely available code for simulated annealing, see http://www.ingber.com/#ASA. In 2007, Georg Holzmann was asking this list if there is an implementation of that powerful code within SciPy, see http://mail.scipy.org/pipermail/scipy-user/2007-November/014549.html . The responses back then were negative, suggesting Georg to wrap it himself and referencing to swig.org. I was wondering if the state of that issue, python wrapper for the ASA code, changed in the last 4.5 years. And if not, is swig.org still the best place to look for starting a wrapper myself? I am aware that there is a Matlab wrapper for ASA, with is +/- up-to-date (Feb 2011), see http://ssakata.sdf.org/software/ . I don't know if this would help to create a wrapper for Python. Many thanks, Dieter On 13/02/12 13:34, Dieter Werthmüller wrote:
Hi there,
The message 'adaptive simulated annealing (ASA)' is over four years old now, asking for a python implementation of ASA (http://www.ingber.com/#ASA).
I was wondering if such an implementation is around today?
Kind regards, Dieter
-- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
![](https://secure.gravatar.com/avatar/ad13088a623822caf74e635a68a55eae.jpg?s=120&d=mm&r=g)
2012/2/21 Dieter Werthmüller <Dieter.Werthmuller@ed.ac.uk>:
Dear reader,
I give it another try with more explanation. If it remains unanswered I will assume that such an implementation does not exist.
No answer and google search doesn't show anything, it looks like "such an implementation does not exist" my impression is global optimization is not very common in this neighborhood. Josef
Adaptive Simulated Annealing (ASA) is a freely available code for simulated annealing, see http://www.ingber.com/#ASA.
In 2007, Georg Holzmann was asking this list if there is an implementation of that powerful code within SciPy, see http://mail.scipy.org/pipermail/scipy-user/2007-November/014549.html .
The responses back then were negative, suggesting Georg to wrap it himself and referencing to swig.org.
I was wondering if the state of that issue, python wrapper for the ASA code, changed in the last 4.5 years. And if not, is swig.org still the best place to look for starting a wrapper myself?
I am aware that there is a Matlab wrapper for ASA, with is +/- up-to-date (Feb 2011), see http://ssakata.sdf.org/software/ . I don't know if this would help to create a wrapper for Python.
Many thanks, Dieter
On 13/02/12 13:34, Dieter Werthmüller wrote:
Hi there,
The message 'adaptive simulated annealing (ASA)' is over four years old now, asking for a python implementation of ASA (http://www.ingber.com/#ASA).
I was wondering if such an implementation is around today?
Kind regards, Dieter
-- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
_______________________________________________ SciPy-User mailing list SciPy-User@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-user
![](https://secure.gravatar.com/avatar/97c543aca1ac7bbcfb5279d0300c8330.jpg?s=120&d=mm&r=g)
On Tue, Feb 21, 2012 at 7:15 PM, <josef.pktd@gmail.com> wrote:
2012/2/21 Dieter Werthmüller <Dieter.Werthmuller@ed.ac.uk>:
Dear reader,
I give it another try with more explanation. If it remains unanswered I will assume that such an implementation does not exist.
No answer and google search doesn't show anything, it looks like "such an implementation does not exist"
my impression is global optimization is not very common in this neighborhood.
Which is to say, probably no-one has gotten around to it (unless Georg Holzmann did and neglected to tell anyone), but I'm sure people would be interested if you were to write and make such code available. SWIG remains a fine way to implement Python wrappers, but I prefer Cython myself. If you go with Cython, the Demos/callback directory in the source tree has a good example of passing a Python function (like a cost function) to a C library: http://cython.org/release/Cython-0.15/Demos/callback/ Another option for is 'ctypes' (which is included with the standard Python distribution, and lets you avoid dealing with C compilers entirely). -- N
![](https://secure.gravatar.com/avatar/b4d84baefad3c955816c604630439500.jpg?s=120&d=mm&r=g)
Hello, On Monday, February 13, 2012 6:34:46 AM UTC-7, Dieter Werthmüller wrote:
The message 'adaptive simulated annealing (ASA)' is over four years old now, asking for a python implementation of ASA (http://www.ingber.com/#ASA ).
I was wondering if such an implementation is around today?
Now there is: http://pypi.python.org/pypi/pyasa Let me know what you think. Robert.
![](https://secure.gravatar.com/avatar/cc747272644252b827b8d734bd6de1da.jpg?s=120&d=mm&r=g)
--- Исходное сообщение --- От кого: "Robert Jordens" <jordens@gmail.com> Кому: scipy-user@googlegroups.com Дата: 13 июля 2012, 18:21:43 Тема: Re: [SciPy-User] [SciPy-user] adaptive simulated annealing (ASA)
Hello,
On Monday, February 13, 2012 6:34:46 AM UTC-7, Dieter Werthmller wrote:
The message 'adaptive simulated annealing (ASA)' is over four years old
now, asking for a python implementation of ASA (http://www.ingber.com/#ASA).
I was wondering if such an implementation is around today? Now there is: http://pypi.python.org/pypi/pyasa
Let me know what you think. I have connected it to OpenOpt GLP class ( http://openopt.org/GLP ); unfortunately, vectorization seems to be unavailable for pyasa, isn't it? Are any possibilities to provide it? Thus currently de and pswarm work many times and sometimes even orders faster. Regards, D.
Robert.
_______________________________________________ SciPy-User mailing listSciPy-User@scipy.orghttp://mail.scipy.org/mailman/listinfo/scipy-user
participants (5)
-
Dieter Werthmüller
-
Dmitrey
-
josef.pktd@gmail.com
-
Nathaniel Smith
-
Robert Jordens