Title: Proximal Gradient Methods with Adaptive Subspace Samplings
Together with Gilles Bareilles, Yassine Laguel, Franck Iutzeler, and Jérôme Malick we propose an asynchronous version of the Progressive Hedging Algorithm(a popular strategy in multistage stochastic programming) that is able to compute an update as soon as a scenario subproblem is solved. Based on the similar arguments as in Asynchronous ADMM, we prove that the asynchronous version has the same convergence properties as the standard one and we release an easy-to-use Julia toolbox .
