Dynamic Ride-Sharing: Theory and Practice

Abstract

Dynamic Ride-sharing (DR) is envisioned as an automated process by which individuals find ride-matches on a trip by trip basis. This paper examines the DR concept on both a theoretical basis and on the basis of actual implementation in Los Angeles. Specifically, the paper investigates the likelihood that the user of a DR system would be successful in finding a ride-match. In a theoretical sense, this paper shows that dynamic ride-sharing is a viable concept. For a congested freeway corridor, the number of trips generated per unit time and space should be sufficient to yield a reasonably large population of potential ride-matches for a DR system. Unfortunately, as demonstrated in the experiment, theory and practice are not the same. Even when individuals share common trip patterns, consummating a ride-match is no easy task for logistical reasons. At best, one might expect a one in five chance of someone offering a ride when trip patterns are similar. This probability would decline for casual trips, and when contacting individuals who have not expressed a prior willingness to carpool.