OSeMOSYS-RDM: A reproducible workflow for robust decision making with OSeMOSYS models

Long-term energy planning under deep uncertainty benefits from stress-testing strategies across many plausible futures. OSeMOSYS-RDM is an open-source, reproducible workflow that couples Robust Decision Making (RDM) with the Open Source energy Modeling System (OSeMOSYS). It automates uncertainty sampling (Latin Hypercube Sampling, LHS), ensemble model execution across multiple solvers, standardized post-processing, and scenario discovery using the Patient Rule Induction Method (PRIM). The workflow is configuration-driven (Excel and YAML), supports workflow reproducibility via Data Version Control (DVC), and produces shareable inputs/outputs for transparent, repeatable studies. Although developed for energy-system models, it can be applied to any OSeMOSYS-encoded system (e.g., land use or industrial processes). This MethodsX article provides a standalone description of the architecture, configuration interface, and outputs to enable reuse and adaptation by the OSeMOSYS community. The workflow automates the quantitative core of an RDM study (experimental design, ensemble execution, and scenario discovery) and is intended to support -not replace- participatory engagement with decision makers through which a full RDM analysis is framed and deliberated.•OSeMOSYS-RDM uses Excel and YAML configuration files to define uncertainties, experiments, and outputs without modifying the core workflow code.•It automates Latin Hypercube Sampling, batch and parallel solver execution (GLPK, CBC, CPLEX, and Gurobi), standardized post-processing into shareable datasets, and DVC-based pipeline execution for reproducible runs.•The method is demonstrated on a multisectoral Uganda model and a Costa Rica transport decarbonization application.