1: \begin{abstract}

2: Copositive optimization is a special case of convex conic programming, and it consists of optimizing a linear function over the cone of all completely positive matrices under  linear constraints. 

3: Copositive optimization provides powerful relaxations of NP-hard quadratic problems or combinatorial problems, but there are still many open problems regarding copositive or completely positive matrices.

4: In this paper, we focus on one such problem; finding a completely positive (CP) factorization for a given completely positive matrix.

5: We treat it as a nonsmooth Riemannian optimization problem, i.e., a minimization problem of a nonsmooth function over a Riemannian manifold.

6: To solve this problem, we present a general smoothing framework for solving nonsmooth Riemannian optimization problems and show convergence to a stationary point of the original problem. 

7: An advantage is that we can implement it quickly with minimal effort by directly using the existing standard smooth Riemannian solvers, such as Manopt.

8: Numerical experiments show the efficiency of our method especially for large-scale CP factorizations.

9: 

10: \smallskip

11: \textbf{Keywords:}

12: completely positive factorization, smoothing method, nonsmooth Riemannian optimization problem

13: 

14: \smallskip

15: \textbf{AMS:}

16: 15A23, 15B48, 90C48, 90C59

17: \end{abstract}

18: