1: \begin{abstract}

2:  We propose a   new

3:  asynchronous parallel block-descent algorithmic framework  for the minimization of the sum of a

4:  smooth nonconvex function and a   {nonsmooth convex}  one, subject to both convex and nonconvex constraints.

5:   The proposed  framework hinges on successive convex approximation techniques and a novel probabilistic model that  captures key elements of  modern computational architectures  and asynchronous implementations in a more faithful way than  current state-of-the-art models.

6:  Other key features of the framework

7:  are: i)  it covers in a unified way several  specific solution methods;  ii) it accommodates a variety of possible parallel computing architectures; and iii) it can deal with nonconvex constraints.

8: Almost sure convergence to stationary solutions  is proved, and theoretical complexity results are provided, showing  nearly ideal linear speedup when the number of workers is not too large.

9: 

10: 

11:  \keywords{ Asynchronous algorithms   \and nonconvex constraints \and  parallel methods \and

12:  probabilistic model.}

13: 

14:   \end{abstract}

15: