Metrics for Quantifying Shareability in Transportation Networks: The Maximum Network Flow Overlap Problem
Cities around the world vary in terms of their transportation networks and travel demand patterns; these variations affect the viability of shared mobility services. This study proposes metrics to quantify the shareability of person-trips in a city, as a function of two inputs--the road network structure and origin-destination (OD) travel demand. The study first conceptualizes a fundamental shareability unit, 'flow overlap'. Flow overlap denotes, for a person-trip traversing a given path, the weighted (by link distance) average number of other trips sharing the links along the original person's path. The study extends this concept to the network level and formulates the Maximum Network Flow Overlap Problem (MNFLOP) to assign all OD trips to paths that maximize network-wide flow overlap. The study utilizes the MNFLOP output to calculate metrics of shareability at various levels of aggregation: person-trip level, OD level, origin or destination level, network level, and link level. The study applies the MNFLOP and associated shareability metrics to different OD demand scenarios in the Sioux Falls network. The computational results verify that (i) MNFLOP assigns person-trips to paths such that flow overlaps significantly increase relative to shortest path assignment, (ii) MNFLOP and its associated shareability metrics can meaningfully differentiate between different OD trip matrices in terms of shareability, and (iii) an MNFLOP-based metric can quantify demand dispersion--a metric of the directionality of demand--in addition to the magnitude of demand, for trips originating or terminating from a single node in the network. The paper also includes an extensive discussion of potential future uses of the MNFLOP and its associated shareability metrics.