Proactive rebalancing and speed-up techniques for on-demand high capacity vehicle pooling.

By proposing speed-up techniques and a proactive rebalance algorithm, we improve the Mobility-on-Demand fleet management approach in [1] on both computational performance and system performance. The speed-up techniques comprise search space pruning and parallelization Input/Output reduction, which reduce the computation time by up to 97.67% in experiments on taxi trips in Manhattan, New York City. The proactive rebalancing algorithm guides idle vehicles to future demand based on a probability distribution estimated using historical data, which increases the service rate by 4.8% on average, and decreases the waiting time and total delay by 5.0% and 10.7% on average respectively.