1: \begin{abstract}

2: We consider algorithms that, from an arbitrarily sampling of $N$ spheres (possibly overlapping), find a close packed configuration without overlapping. 

3: These problems can be formulated as minimization problems with non-convex constraints. 

4: For such packing problems, we observe that the classical iterative Arrow-Hurwicz algorithm does not converge.

5: We derive a novel algorithm from a multi-step variant of the Arrow-Hurwicz scheme with damping.

6: We compare this algorithm with classical algorithms belonging to the class of linearly constrained Lagrangian methods and show that it performs better.

7: We provide an analysis of the convergence of these algorithms in the simple case of two spheres in one spatial dimension. 

8: Finally, we investigate the behaviour of our algorithm when the number of spheres is large.

9: 

10: \end{abstract}

11: