When the uncertainty is explicitly modeled in an optimization problem, it is often necessary to use samples to compute a solution, which gives rise to a need to compute confidence intervals around the objective function value that is obtained. In this paper we describe software that implements well-known methods for two stage problems and we provide extensions to multiple stages. The software is available as part of the mpi-sppy package Knueven et al. (2021), which is available for download from https://github.com/Pyomo/mpi-sppy.
Article
View Confidence Interval Software for Multi-stage Stochastic Programs