Riemannian linearized proximal algorithms for nonnegative inverse eigenvalue problem (Q6141535)

The authors formulate the nonnegative inverse eigenvalue problem as a convex composite optimization problem on Riemannian manifolds. A scheme of the Riemannian linearized proximal algorithm is proposed to solve the problem. The local and global convergence of the proposed method is established under the assumptions of local weak sharp minima order and a quasi-regularity condition. The numerical performance of the proposed method for large scale problems and sparse matrices is presented.