Abstract
A consequence of rapid growth in urban population is the increasing demand for urban freight, which is transported primarily by commercial vehicles. Increasing the transport infrastructure, including overnight parking supply, to accommodate the growing population of commercial vehicles in land-scarce cities is a challenge. Lack of understanding of the parking behavior of commercial vehicles and absence of tools for policy analysis makes the task onerous for the urban planner. With data collected from Singapore, we develop an operational discrete choice model to explain the overnight parking behavior of commercial vehicles. We encounter missing values and omitted attributes in the data, which, if ignored, result in biased parameter estimates. To overcome the twin data deficiencies while estimating discrete-choice models, a novel algorithm combining the methods of multiple imputation and control function is proposed. We validate the effectiveness of the algorithm in a Monte-Carlo simulation and demonstrate its application in correcting price endogeneity in the overnight parking choice model. Development of a robust parking policy for commercial vehicles requires analysis of multiple scenarios. Existing urban traffic simulators lack capabilities to perform analysis of overnight parking policies. To aid the urban planner in evaluating parking policies, we enhance the capabilities in SimMobility, an agent-based urban passenger and freight traffic simulation platform and demonstrate an application in a hypothetical scenario, where the existing parking supply is relocated closer to industrial areas in Singapore. While the redistribution of parking capacity fulfills the objective of releasing land parcels in residential and commercial districts for other potential uses, it has an adverse impact on vehicle operations by increasing the length of empty trips. We are also able to quantify the changes in heavy vehicle traffic on the road network.