Stochastic programming problems model future uncertainty through the analysis of possible future outcomes (scenarios): the more detailed the description is, the more robust are the decisions taken.
Real life applications may require the generation of very large scenario trees and, consequently, of large-scale size deterministic equivalent matrices. In a large scenario tree, the large-scale problem can provide a fine-grained solution to a problem that could have been solved more coarsely by using a much smaller tree. This observation suggests an idea for a warm-start technique that can be applied in the context of interior point methods.
In this talk we explore how to obtain a warm-start point by solving the stochastic optimization problem for a reduced event tree and then use it as an advanced iterate for the complete formulation.
Current 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996