The qutip-qoc package builds up on the qutip-qtrl package.
It enhances it by providing two additional algorithms to optimize analytically defined control functions. The first one is an extension of Gradient Optimization of Analytic conTrols (GOAT). The second one (JOPT) leverages QuTiPs version 5 new diffrax abilities to directly calculate gradients of JAX defined control functions using automatic differentiation.
Both algorithms consist of a two-layer approach to find global optimal values for parameterized analytical control functions.
The global optimization layer provides scipy.optimize.dual_annealing and scipy.optimize.basinhopping, while the local minimization layer supports all
gradient driven scipy.optimize.minimize methods.
The package also aims for a more general way of defining control problems with QuTiP and makes switching between the four control algorithms (GOAT, JOPT, and GRAPE and CRAB implemented in qutip-qtrl) very easy.
As with qutip-qtrl, the qutip-qoc package aims at providing advanced tools for the optimal control of quantum devices.
Compared to other libraries for quantum optimal control, qutip-qoc puts additional emphasis on the physics layer and the interaction with the QuTiP package.
Along with the extended GOAT and JOPT algorithms the package offers support for both the CRAB and GRAPE methods defined in qutip-qtrl.
If you would like to know the future development plan and ideas, have a look at the qutip roadmap and ideas.
To install the package, use
pip install qutip-qoc
The documentation of qutip-qoc updated to the latest development version is hosted at qutip-qoc.readthedocs.io.
Tutorials related to using quantum optimal control in qutip-qoc can be found here.
If you want to edit the source code, please download the source code and run the following command under the root qutip-qoc folder,
pip install --upgrade pip
pip install -e .
which makes sure that you are up to date with the latest pip version. Contribution guidelines are available here.
To build and test the documentation, additional packages need to be installed:
pip install pytest matplotlib sphinx sphinxcontrib-bibtex numpydoc sphinx_rtd_theme sphinxcontrib-bibtex
Under the doc directory, use
make html
to build the documentation, or
make doctest
to test the code in the documentation.
To test the installation, choose the correct branch that matches with the version, e.g., qutip-qoc-0.2.X for version 0.2. Then download the source code and run from the qutip-qoc directory.
pytest tests
If you use qutip-qoc in your research, please cite the original QuTiP papers that are available here.
This package is supported and maintained by the same developers group as QuTiP.
QuTiP development is supported by Nori's lab at RIKEN, by the University of Sherbrooke, by Chalmers University of Technology, by Macquarie University and by Aberystwyth University, among other supporting organizations.
You are free to use this software, with or without modification, provided that the conditions listed in the LICENSE.txt file are satisfied.