Rebalancing Docked Bicycle Sharing System with Approximate Dynamic Programming and Reinforcement Learning

Figure 4

Computation time needed for 10 iterations.