Performance Comparison of Supply-Demand Matching Policies for On-Demand Mobility Services

In the context of on-demand mobility services, we compare the performance of three matching policies of increasing sophistication on scenarios based on real data from the city of Chicago in the United States. The comparative study examines the influence of prebooking and ridesharing on the gap between the different policies. We find that more sophisticated approaches can improve the acceptance ratio and the service-level key performance indicators at the expense of longer computation times. Prebooking appears to consistently give an edge to more sophisticated policies by providing advance information and thus the flexibility to make better plans. The effect of ridesharing is less straightforward to isolate. But, again, prebooking helps more sophisticated approaches reduce excess ride time, a direct consequence of ridesharing.