Goods movement accounts for a significant and growing share of urban traffic, energy use and greenhouse gas emissions (GHGs). This project investigated the vehicle miles travelled (VMT) and emissions impact of on-demand food delivery under different COVID-19 pandemic periods and multiple operational strategies, with real-world scenarios set up in the city of Riverside, California. The evaluation results showed that during COVID-19 the total VMT and pollutant emissions (CO2, CO, HC, NOx) incurred by eat out demand all decreased by 25% compared with the before-COVID-19 period. The system can achieve substantial reductions in vehicle trips and emissions with higher penetration of on-demand delivery. From the dynamic operation perspective, the multi-restaurant strategy (allow food orders to be bundled from multiple restaurants in one driver’s tour) can bring 28% of VMT and emissions reductions while avoiding introducing additional delay compared to the one-restaurant policy (only allow food orders from the same restaurant to be bundled in one driver’s tour). The research results indicate that the delivery platform should provide more reliable service with lower cost to increase the food delivery penetration level, which needs improvement in driver capacity management, eco-friendly delivery strategy, and efficient order allocation system. Meanwhile, to achieve maximum VMT and emissions reduction, the platform should encourage order bundling and employ a multi-restaurant policy to provide higher flexibility to group food orders, especially from restaurants located densely in one shopping plaza or commercial zone.
