Paper Title
Applying Monte Carlo Tree Search to the Optimization of Train Speed Curve for Energy-Saving in a Mass Rapid Transit System
Abstract
This study applied combinatorial optimization techniques to Monte Carlo tree search (MCTS) and incorporated the
calculation of the optimal speed curve for energy-saving mass rapid train operation. The goal was to achieve online speed
command planning based on this derived curve when the trains operate between successive stations. MCTS, an algorithm
originally designed for two-player games, was subsequently adjusted and transformed into an adequate optimization method.
In addition, energy-saving factors, coasting and regenerative braking were considered in the experiments, the results of which
confirmed that MCTS can indeed solve the problem of energy-saving train operation.
Index Terms - MCTS, MRTS, Train Speed, Energy-Saving