Topological properties of ride-pooling shareability graph as a proxy to key performance indicators

Abstract

Ride-pooling is a promising and actively developing branch of urban mobility. Algorithmic solutions match travellers with feasible rides using a so-called shareability graphs. However, analysis in the field present very little information on the topological structure of such graphs. In the paper, we experiment with varying demand levels and behavioural parameters of travellers sharing rides in Delft, Netherlands, to observe changes in the graph structure and see how properties of the graph correlate with ride-pooling KPIs. We observed strong relations in our experimental study, suggesting that there is potential to use Network Science tools in order to better understand shareability in the ride-pooling problem.