Understanding The Impacts of On-Demand Delivery of Prepared Meals and Grocery

摘要: The growing technology access and effects of social distancing policies during the COVID-19 pandemic significantly increased the demand for online purchases. Understanding the potential mobility impacts of e-commerce operations at a regional level becomes relevant. This study uses data from multiple sources to develop a modeling framework for on-demand delivery (ODD) operations for meals and groceries in POLARIS, a large-scale activity-based travel demand model and mesoscopic traffic simulator. Deliveries are simulated following similar characteristics of taxi operations, allowing for orders to be picked up by individuals on personal vehicles. The demand is modeled from a behavioral survey using a zero-one inflated Beta model aiming to address the excess of zeros in the data and the presence of ones. Estimations for low and high e-commerce demand are analyzed for operations in Austin, Texas, from 2020 to 2040 as the population grows by 60%. It is expected that ODD operations will increase in future years, leading to a significant increment in operational vehicle miles traveled (VMT). Although substitutionary trips come from exchanging a percentage of driving activities for e-commerce operations, a significant part of delivery operations corresponds to empty miles with driver repositioning or idling. Additional analysis of the effect of delivery policies for short trips indicates that nearly half of the VMT can be done using low-emission vehicles like bicycles. This study highlights the importance of modeling ODD operations in transportation system simulations analysis due to the high impact that these operations can have in the long term.