Designing efficient shared mobility services directly adapted tothe spatio-temporal features of the potential demand

Abstract

2 Shared mobility services are announced as a game-changer in transportation and a promising so3 lution to reduce congestion and improve the performance of urban mobility. They could prefigure 4 the arrival of autonomous vehicles. Modeling of these new services is a real challenge, especially 5 because existing approaches are mainly an adaptation of methods devoted to classic transportation 6 services. Consequently, this paper introduces a new data-driven optimization method fully devoted 7 to shared mobility service. First, the proposed approach decomposes the recurrent demand based 8 on its spatio-temporal features to overcome the drawbacks of the existing methods. Notably, it 9 makes it possible to consider larger instances and to build robust solutions. Thus, recurrent de10 mand patterns are identified to capture the potential demand of shared mobility services using a 11 tailored clustering process. Second, a variant of Dial-a-Ride Problem is implemented to design 12 robust lines to serve this demand. Such a hybrid method makes it possible to define relatively 13 massive transport lines while maintaining spatial and temporal proximity to users real demand. 14 The method is then tested with an open-source dataset released by the New York City Taxi and 15 Limousine Commission.

16 Keywords: Clustering, Shared mobility, Dial-a-Ride, Similarity, Ride-sharing.

1 List of Figures 2 1 (a),(b),(c),(d) clusters with different characteristics, the pick-up are depicted in green and drop off 3 in red. nk denotes the number of trips in the cluster k, lk denotes the average length of trips in k and 4 τk denotes the average duration of trips in k. (e) Ratio of clustered trips per day in Midtown and 5 Upper East Side from 8h to 11h. (f) Example of demand graph for a randomly selected meta-cluster. 7 6 2 Green markers depict the pick-up and red markers the drop off (a) 3 meta-clusters randomly chosen 7 between 08h15 and 08h35. (b) example of line designed, serving the centroids of pick-up and drop 8 off of each meta-cluster. (c) Total number of trips per day served on the set of 3 meta-clusters. (d) 9 Shows the number of meta-clusters for each time slot in function of the minimal median value of 10 trips per day required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 11 3 (a) depicts the total number of users effectively served in the set of meta-clusters selected in Section 12 - Selection of the spatio-temporal areas. (b),(c),(d) shows for each period presented in Table 2, the 13 customized lines found. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

14 List of Tables 15 1 Average spatial and temporal distances between pick-up, drop off and the centroïds 16 of the meta-clusters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 17 2 Result of the search of rounds for the 3 time periods from 08h00 to 11h00 . . . . . 13