Processed data and evidence for Evo-MOMBRL bridge-network maintenance experiments

Curated model inputs and processed evidence accompanying the article “Evo-MOMBRL: Evolutionary multi-objective model-based reinforcement learning for long-term bridge network maintenance”. The archive contains sanitized asset inputs, full-precision transition matrices, pairwise road-network consequence inputs, exact three- and four-bridge validation, five-run learning-method fronts, paired ablations, exhaustive CBM/TBM results, and aggregate engineering trajectories. Raw upstream NBI/TIGER downloads, model checkpoints, internal logs, and superseded experiments are excluded.