Uncertainty-Aware Routing in Delay-Tolerant Networks: Artifact and Reproducibility Package

This artifact provides a curated experimental framework for studying routing in Delay-Tolerant Networks (DTNs) under uncertain contact plans, enabling reproducible evaluation across diverse modeling assumptions and decision paradigms. Organized around a five-dimensional taxonomy of uncertainty-aware routing, the repository includes reference implementations spanning optimal Markov Decision Process formulations (RUCoP), statistical model checking (LSS), reinforcement learning approaches (QL and GROGU), and belief-based planning models such as POMDP routing under dependent node failures (DNF) and replication-aware decision making (RAD). The artifact further offers standalone analytical examples in Python and Julia, realistic satellite contact plans, and integrated tooling—including the MODEST Toolset and the POMDPs.jl ecosystem—to support experimentation with both centralized and distributed policies. Developed as Deliverable D3.3 of the MISSION project on model-based engineering of dependable space systems, this release promotes transparent comparison of routing strategies and serves as a research platform for advancing uncertainty-aware communication in future space and intermittently connected networks.