Evaluating the Impact of Urban Consolidation Center and Off-Hour Deliveries on Freight Flows to a Retail District Using Agent-Based Simulation

Abstract

With rapid ongoing urbanization, cities across the world face a multitude of challenges in urban logistics. Delivery of goods to retail districts is particularly challenging as these places are typically located in congested urban centers. In response, policy makers have explored various freight management initiatives, including urban consolidation centers (UCC) and off-hour deliveries (OHD). This study examines the impact of these initiatives on freight flows to a retail district in Singapore. The study approach pairs empirical behavioral models and an agent-based simulation. First, using results from a stated preference survey, the choice behavior of two relevant actors—establishments that ship goods and establishments that receive goods within a retail district—and their likelihood of participating in UCC or OHD are analyzed. Then, the resulting behavioral models are incorporated into a city-scale agent-based simulator to evaluate the impact of these initiatives on freight flows, tracking multiple logistics-related performance indicators. The results show that the likelihood of participating in UCC and OHD declines when multiple actors are involved in the participation decision. Both UCC and OHD have the potential to reduce freight traffic and parking demand, although in different ways. For UCC, a minimum level of participation must be achieved to guarantee increase in vehicle load factors. OHD decreases load factors as well as the required number of dedicated trips to the retail district.

How artificial intelligence cooperating with agent‐based modeling for urban studies: A systematic review

“…Xiao and Lo (2016) developed a social network‐based framework of departure time choices using Bayesian learning and applied it to ABM to evaluate its effectiveness. In addition, some other ML methods can be used to simulate agents' attitudes (Xu et al., 2020), action execution or not (Yao et al., 2020), the probability to act (Mepparambath et al., 2023; Zhao & Peng, 2015), the tendency (Li et al., 2020), mental representations (Kusumastuti et al., 2010), etc., which can trigger agents' the next behaviors when they meet some conditions or are agreed by the policymaking agents.…”