Reinforcement learning for bus dispatch: an LSPI agent (LSTD-Q) decides how many buses to run every 15 min on a simulated 12-stop corridor, cutting daily cost 30% vs a fixed timetable. Value function audited with LSTD(λ). Interactive Flask + Chart.js dashboard.
python machine-learning reinforcement-learning simulation numpy transportation chartjs data-visualization least-squares operations-research policy-iteration smart-city demand-management lstd temporal-difference-learning markov-decision-process public-transit value-function-approximation lspi
-
Updated
Nov 15, 2026 - Python