Riemannian Manifold Optimization Library

This package provides many state-of-the-art algorithms to optimize a smooth cost function defined on a Riemannian manifold. The package is written in C++ and uses the standard linear algebra libraries: BLAS and LAPACK. It can be used alone in a C++ environment or in Matlab with a Mex interface. The package is more reliable and requires smaller computational time compared with code written only in Matlab. Users need only provide a cost function, gradient function, and the action of the Riemannian Hessian (if a Newton method is used) in Matlab or C++. The package optimizes the function given a set of user-specified parameters, e.g., the domain manifold, algorithm, stopping criterion.