This paper presents a Python package to solve multi-stage stochastic linear programs (MSLP) and multi-stage stochastic integer programs (MSIP). Algorithms based on an extensive formulation and Stochastic Dual Dynamic (Integer) Programming (SDDP/SDDiP) method are implemented. The package is synthetically friendly and has a number of features which are not available in the competing software packages. In particular, the package deals with some of the restrictions on the underlying data process imposed by the previously available software packages. As an application of the package, three large-scale real-world problems - power system planning, portfolio optimization, airline revenue management, are discussed.
View A Python package for multi-stage stochastic programming