Rapid Optimization Library (ROL)

ROL (as in rock and roll) is a high-performance C++ library for numerical optimization. ROL brings an extensive collection of state-of-the-art optimization algorithms to virtually any application. Its programming interface supports any computational hardware, including heterogeneous many-core systems with digital and analog accelerators. ROL has been used with great success for optimal control, optimal design, inverse problems, image processing and mesh optimization, in application areas including geophysics, structural dynamics, fluid dynamics, electromagnetics, quantum computing, hypersonics and geospatial imaging.

For additional details, see https://rol.sandia.gov.

Feature highlights:

1. Vector abstractions and matrix-free interface for universal applicability
2. Modern algorithms for unconstrained and constrained optimization
3. Easy-to-use methods for stochastic and risk-aware optimization
4. Fast and robust algorithms for nonsmooth optimization
5. Trust-region methods for inexact and adaptive computations
6. PDE-OPT application development kit for PDE-constrained optimization
7. Interfaces and algorithms for optimal experimental design

Getting Started

ROL is a C++ library with a cmake build system. There are minimal third-party library requirements consisting of BLAS and LAPACK which also require a suitable Fortran compiler on the system. On Linux-based systems with an apt package manager, the dependencies may be installed with

apt install cmake gfortran libopenblas-dev liblapack3

Some typical configure scripts may be found in the .github/workflows/ directory. An in-source release build including ROL's examples and tests may be configured from within the source directory with

cmake -D CMAKE_BUILD_TYPE:STRING=RELEASE \
      -D ENABLE_EXAMPLES:BOOL=ON \
      -D ENABLE_TESTS:BOOL=ON \
      -B build \
      .

After a successful configure, ROL is built by then changing to the build directory and running make

cd build
make

Copyright and License

See COPYRIGHT and LICENSE.

Questions?

Contact team or developers:

• ROL Team (GitHub handle: @sandialabs/rol)
• Drew Kouri (GitHub handle: dpkouri or [email protected])
• Denis Ridzal (GitHub handle: dridzal or [email protected])