pySOT is an asynchronous parallel optimization toolbox for global deterministic optimization problems. The main purpose of the toolbox is for optimization of computationally expensive black-box objective functions with continuous and/or integer variables where the number of evaluations is limited. If there are several processors available it may make sense to evaluate the objective function using either asynchronous or synchronous parallel. pySOT uses the event-driven framework for asynchronous optimization strategies POAP (https://github.com/dbindel/POAP) to provide this functionality.
The toolbox is hosted on GitHub: https://github.com/dme65/pySOT
Link to the pySOT documentation: https://github.com/dme65/pySOT/blob/master/docs/pySOT.pdf
pySOT has been downloaded 12,492 times from 2015-June-4 to 2016-January-24
Make sure you have Python 2.7.x and pip installed. The easiest way to install pySOT is using:
pip install pySOT
Several pySOT examples can be found at: https://github.com/dme65/pySOT/tree/master/pySOT/test
Q: I can't find the GUI
A: You need to install PySide
Q: I can't find the MARS interpolant
A: You need to install py-earth (https://github.com/jcrudy/py-earth)
Q: I used pySOT for my research and want to cite it
A: There is currently no published paper on pySOT so we recommend citing pySOT like this: D. Eriksson, D. Bindel, and C. Shoemaker. Surrogate Optimization Toolbox (pySOT). github.com/dme65/pySOT, 2015
Q: Is there support for Python 3?
A: pySOT currently doesn't support Python 3, mainly because of some pySOT dependencies lacking Python 3 support.