### Anthony Papavasiliou (Université Catholique de Louvain, Belgium)

#### An asynchronous distributed algorithm for solving stochastic unit commitment

*Wednesday 26 October 2016 at 15.00, JCMB 5215*

##### Abstract

We present an asynchronous algorithm for solving the stochastic unit
commitment (SUC) problem using scenario decomposition. The algorithm is
motivated by the scale of problem and significant differences in run times
observed among scenario subproblems, which can result in an inefficient use of
distributed computing resources by parallel synchronous algorithms. Iterations
are performed asynchronously using a block-coordinate subgradient method which
allows performing block-coordinate updates using delayed information. We
provide convergence guarantees for the asynchronous block-coordinate
subgradient descent method based on previous results for incremental
subgradient methods and stochastic subgradient methods. The algorithm recovers
candidate primal solutions from the solutions of scenario subproblems using
re-combination heuristics in parallel to dual iterations.

The asynchronous algorithm is implemented in a high performance computing
cluster and we conduct numerical experiments for two-stage SUC instances of
the Western Electricity Coordinating Council system (WECC, hourly resolution,
130 thermal generators, 182 nodes and 319 lines) with up to 1000 scenarios and
of the Central Western European system (CWE, 15-minute resolution, 656 thermal
generators, 679 nodes and 1073 lines) with up to 120 scenarios. While using 10
nodes of the cluster per instance, the algorithm is able to provide 2%
suboptimal solutions to all problems within operationally acceptable time
frames: 47 minutes for WECC and 3 hours, 54 minutes for CWE. Moreover, we find
that an equivalent synchronous parallel subgradient algorithm would leave
processors idle between 25% and 84% of the time, which further stresses the
need for designing asynchronous optimization schemes in order to fully exploit
distributed computing.

### Seminars by year

*Current*
*2016*
*2015*
*2014*
*2013*
*2012*
*2011*
*2010*
*2009*
*2008*
*2007*
*2006*
*2005*
*2004*
*2003*
*2002*
*2001*
*2000*
*1999*
*1998*
*1997*
*1996*