As rideshare and e-hailing services grow in popularity as a regular feature of daily commutes, their impact on traffic congestion, commuting cost, and travel behavior becomes increasingly significant. With the emergence of a range of companies, most notably Uber and Lyft, providing these services at a variety of price points, customers have greater flexibility in travelling from one point to another. Despite the growing need to study the specific impacts of these new modes of transportation, planners and policymakers have access to very little real data to support policy decisions. Most existing research focuses solely on one aspect of travel, the morning commute, which may not be an accurate representation of information like the number of drivers and total vehicle miles traveled (VMT). These metrics provide important information about commutes such as whether these services are negatively impacting road congestion.

Given the existing research gap and the uncertainty of the impacts of rideshare/e-hailing on equity, the environment, and efficiency, a research project entitled “Modeling e-hailing and car-pooling services in a coupled morning-evening commute framework” was funded by the Pacific Southwest Region University Transportation Center (PSR). Researchers proposed a ridesharing model to improve system performance on gathering metrics related to the number of drivers, VMT, etc. The research also examined how such a model can help achieve social and environmental benefits by identifying some feasible and flexible commute options for all commuters as well as incentives to reduce VMT. Reducing VMT reduces traffic congestion, thereby, lessening traffic emissions. The research team was a unique collaboration between University of California, Davis and the University of Southern California, including USC Professor of Industrial and Systems Engineering Jong-Shi Pang; UC Davis Professor of Civil and Environmental Engineering Michael H. Zhang; USC Professor of Industrial and Systems Engineering Maged M. Dessouky; and USC doctoral student Wei Gu. This study provided an often-rare opportunity to benefit from a more diverse set of resources and collaborative work, made possible by the PSR research network

The researchers developed a parallel comparison between two models—one coupled and the other decoupled. The decoupled model studies the morning and evening time periods separately. The decoupled model exaggerates commuting demands by counting each individual commuting in the morning/evening as separate riders, which does not account for a rider’s choice in alternating between different transportation modes. For instance, simply assuming that a ride hail trip to work means a ride hail trip from work will overestimate ride hail demand. A commuter may ride hail to work but share a ride from work with a co-worker or choose to take public transportation. Because of these limitations, the research team sought to develop an alternative. In the coupled model, rideshare/e-hailing commutes that occur in the morning and commutes that occur in the afternoon are studied together, providing a holistic understanding of rideshare/e-hailing commuting options. This allows researchers to more accurately register the number of riders, as well as the number of drivers needed to satisfy demand.

Using simulation modeling, the researchers found that the coupled model is more effective in capturing the mode switches between morning and evening, compared with a decoupled morning/evening commute mode. Specifically, the coupled model produces 24.2% fewer drivers and 8.4% less VMT in the system compared with the decoupled model. The decoupled model may not capture costs related to ridesharing inconveniences in the evening. The coupled model enables policymakers and transportation planners to better understand rider/driver/company interactions throughout the day, rather than at one particular time period. Understanding temporal patterns can help guide decisions on promoting incentives to increase vehicle occupancy to reduce congestion and vehicle miles traveled.

Stakeholders such as service companies attempting to predict the number of drivers needed to satisfy consumer demand, residents concerned about congestion and the availability of services, as well as local governments trying to regulate this new, emergent form of transportation, attempt to maximize the benefits of ridesharing/e-hailing while mitigating the potential consequences of their expanded use. This study shows that the highly interconnected relationship between rideshare/e-hailing companies, drivers, and riders not only affect each other, but also the broader system of travel. Models like the coupled model will be critical in developing a more comprehensive transportation plan, as well as enabling better predictive models for commuter traffic. They will be useful in allowing cities to identify which areas to target for ridesharing programs and shuttle services or to assess the impact of policies, such as the establishment of fees on single person ride hail trips. Rapidly changing urban mobility services, especially related to the development of new services, complicate interactions between drivers, riders, and servicing companies. Models and frameworks like this will help to simplify complex problems and make them more understandable to policy makers.

About the Authors:

Patricia (Georgie) Suico is an undergraduate studying Political Science in the CSULB College of Liberal Arts. She works as a staff writer for the METRANS News team.

Kieu-Anh Vu is an undergraduate student studying International Business in the CSULB College of Business. She works as a staff writer for the METRANS News team.