Search for: All records

Trip generation, a critical first step in travel demand forecasting, requires not only estimating trips from the observed sample data, but also calculating the total number of trips in the population, including both the observed trips and the trips missed from the sample (we call them missing trips in this paper). The latter, how to recover missing trips, is scarcely studied in the academic literature, and the state-of-the-art practice is through the application of sample weights to extrapolate from observed trips to the population total. In recent years, big location-based service (LBS) has become a promising alternative data source (in addition to household travel survey data) in trip generation. Because users self-select into using different mobile services that result in LBS data, selection bias exists in the LBS data, and the kinds of trips excluded or included differ systematically among data sources. This study addresses this issue and develops a behaviorally informed approach to quantify the selection biases and recover missing trips. The key idea is that because biases reflected in different data sources are likely different, the integration of multiple biased data sources will mitigate biases. This is achieved by formulating a capture probability that specifies the probability of capturing a trip in a data set as a function of various behavioral factors (e.g., socio-demographics and area-related factors) and estimating the associated parameters through maximum likelihood or Bayesian methods. This approach is evaluated through experimental studies that test the effects of data and model uncertainty on its ability of recovering missing trips. The model is also applied to two real-world case studies: one using the 2017 National Household Travel Survey data and the other using two LBS data sets. Our results demonstrate the robustness of the model in recovering missing trips, even when the analyst completely mis-specifies the underlying trip generation process and the capture probability functions (for quantifying selection biases). The developed methodology can be scalable to any number of data sets and is applicable to both big and small data sets. History: This paper has been accepted for the Transportation Science Special Issue on Machine Learning Methods for Urban Mobility. Funding: This work was supported by the Division of Civil, Mechanical and Manufacturing Innovation [Grant 2114260], the National Institute of General Medical Sciences [Grant 1R01GM108731-01A1], and the U.S. Department of Transportation [Grant 69A3551747116]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/trsc.2024.0550 . 

Policymakers must make management decisions despite incomplete knowledge and conflicting model projections. Little guidance exists for the rapid, representative, and unbiased collection of policy-relevant scientific input from independent modeling teams. Integrating approaches from decision analysis, expert judgment, and model aggregation, we convened multiple modeling teams to evaluate COVID-19 reopening strategies for a mid-sized United States county early in the pandemic. Projections from seventeen distinct models were inconsistent in magnitude but highly consistent in ranking interventions. The 6-mo-ahead aggregate projections were well in line with observed outbreaks in mid-sized US counties. The aggregate results showed that up to half the population could be infected with full workplace reopening, while workplace restrictions reduced median cumulative infections by 82%. Rankings of interventions were consistent across public health objectives, but there was a strong trade-off between public health outcomes and duration of workplace closures, and no win-win intermediate reopening strategies were identified. Between-model variation was high; the aggregate results thus provide valuable risk quantification for decision making. This approach can be applied to the evaluation of management interventions in any setting where models are used to inform decision making. This case study demonstrated the utility of our approach and was one of several multimodel efforts that laid the groundwork for the COVID-19 Scenario Modeling Hub, which has provided multiple rounds of real-time scenario projections for situational awareness and decision making to the Centers for Disease Control and Prevention since December 2020.