Group for Research in Decision Analysis

Interior Point Warmstarts and Stochastic Programming

Andreas Grothey University of Edinburgh, United Kingdom

I present progress on an interior point-based multi-step solution approach for stochastic programming problems. Our approach works with a series of scenario trees that can be seen as successively more accurate discretizations of an underlying probability distribution and employs interior-point method warmstarts to "lift" approximate solutions from one tree to the next larger tree. We give theoretical and computational results that show significant improvements on solution times on a variety of two- and multi-stage stochastic programming problems.