Project adress: https://github.com/pierre-haessig/python-dmpc
dmpc is simulation tool for Model Predictive Control (MPC) and Distributed MPC,
written in pure Python.
As of now, it is in a very early stage, meaning that only a few subset of features are implemented (one type of MPC). However, what is implemented should work well enough and be covered by a resonable set of tests.
Beyond the usual stack of scientific Python packages (in fact only numpy),
dmpc requires the optimization package cvxopt (http://cvxopt.org/)
to solve linear or quadratic optimization problems.
To install dmpc directly from the source tree:
pip install git+git://github.com/pierre-haessig/python-dmpc.git
or clone the repository first and then run:
pip install .
Some examples are provided in the source tree.
As of now, only a heating example with a quadratic cost for the temperature deviations is implemented.
