Smart Transportation Without Neurons -- Fair Metro Network Expansion with Tabular Reinforcement Learning

In three linesTabular reinforcement learning for metro network expansion (MNEP). Reformulated as Non-Markovian Rewards Decision Process (NMRDP): matches Deep RL performance with 18× fewer training episodes and 12× lower carbon emissions. Incorporates social equity criteria. Validated on Xi'an and Amsterdam.

Summary generated by Claude — human-verified