Y. Leng, L. Rudolph, A.S. Pentland, J. Zhao, H.N. Koutsopolous: "Managing travel demand: Location recommendation for system eciency based on mobile phone data". 2016. Proceedings of Data for Good Exchange (D4GX) 2016. New York, NY.

Abstract

Growth in leisure travel has become increasingly signicant economically, socially, and environmentally. However, flexible but uncoordinated travel behaviors exacerbate trafife congestion. Mobile phone records not only reveal human mobility patterns, but also enable us to manage travel demand for system efficiency. In this paper, we propose a location recommendation system that infers personal preferences while accounting for constraints imposed by road capacity in order to manage travel demand. We first infer unobserved preferences using a machine learning technique from phone records. We then formulate an optimization method to improve system Fe ciency. Coupling mobile phone data with trafife counts and road network infrastructures collected in Andorra, this study shows that uncoordinated travel behaviors lead to longer average travel delay, implying the opportunities in managing travel demand by collective decisions. The interplay between congestion relief and overall satisfied location preferences observed in extensive simulations indicate that moderate sacrifices of individual utility lead to significant travel time savings. Specifically, the results show that under full compliance rate, travel delay fell by 52% at a cost of 31% less satisfaction. Under 60% compliance rate, 41% travel delay is saved with a 17% reduction in satisfaction.This paper highlights the effectiveness of the synergy among collective behaviors in increasing system efficiency.