This framework provides high-level abstraction to compute gradient, hessians and tangent linear models for PDE-constrained optimization problems with ODE constraints. It is based on PETSc, the Portable, Extensible Toolkit for Scientific Computation. The FEM capability is based on the C++ library MFEM.
First install PETSc,
and then set appropriate values for PETSC_DIR and PETSC_ARCH in your
environment (PETSc version greater or equal 3.11 is required):
$ export PETSC_DIR=/home/user/petsc $ export PETSC_ARCH=arch-linux2-c-debug
Clone the Git repository hosted at Github
$ git clone https://github.com/stefanozampini/petscopt.git
Finally, enter PETScOPT top level directory and use make to compile
the code and build the PETScOPT library:
$ cd petscopt $ make
Example codes are located at src/ts/examples/tests and src/tao/examples/tests.
In order to enable the MFEM based FEM layer (MFEM version greater or equal 4.3 must be used), you have two options
recommended: Configure PETSc using --download-mfem --download-mfem-commit=v4.3 --download-hypre --download-metis --with-cxx-dialect=C++11
harder: After having built PETSc, follow the instructions to build MFEM, and enable PETSc as a third-party package via
MFEM_USE_PETSC=YES. Then set the environment variableMFEM_DIR:$ export MFEM_DIR=_location_of_MFEM_
Then, the PETScOPT library can be built as:
$ cd petscopt $ make distclean $ make config with_mfem=1 $ make
Example codes are located at src/mfemopt/examples/tests.
In order to check the successfull build of the library type:
$ make check
The PETScOPT library uses the PETSc testing infrastructure. To run the complete test suite type:
$ make test
This project was supported by the Extreme Computing Research Center Division of Computer, Electrical, and Mathematical Sciences & Engineering (CEMSE), King Abdullah University of Science and Technology (KAUST).