Search for: All records

This paper seeks to improve an underutilized conventional bus route by converting it into a semiflexible transit system where passengers provide advance notice of their intended stops, allowing buses to skip downstream stops without demand by taking shortcuts. This approach increases stop density, reduces walking distances to and from bus stops, and maintains operational efficiency. To design this system, we develop optimization models that maximize the number of stops while adhering to tour duration and arrival time constraints. A case study in Allegany County, Maryland, demonstrates significant enhancements for routes that were both underutilized (where the probability of a stop lacking demand exceeded 45%) and had layouts conducive to substantial shortcuts. In these instances, the number of stops can be increased by up to 160%, with the actual improvement depending on route configuration, passenger demand, and advance notice requirements. Funding: Financial support from the the National Science Foundation [Grant 2055347] is gratefully acknowledged. Supplemental Material: The online appendix is available at https://doi.org/10.1287/trsc.2024.0561 . 

Individual evacuation decision making has been studied for multiple decades mainly using theory-based approaches, such as random utility theory. This study aims to bridge the research gap that no studies have adopted data-driven approaches in modeling the compliance of hurricane evacuees with government-issued evacuation orders using survey data. To achieve this, we conducted a survey in two coastal metropolitan regions of Florida (Jacksonville and Tampa) during the 2020 Atlantic hurricane season. After preprocessing survey data, we employed three supervised learning algorithms with different complexities, namely, multinomial logistic regression, random forest, and support vector classifier, to predict evacuation decisions under various hypothetical hurricane threats. We found that the evacuation decision is mainly determined by people’s perception of hurricane risk regardless of whether the government issued an order; COVID-19 risk is not a major factor in evacuation decisions but influences the destination type choice if an evacuation decision is made. Additionally, past and future evacuation destination types were found to be highly correlated. After comparing the algorithms for predicting evacuation decisions, we found that random forest can achieve satisfactory classification performance, especially for certain categories or when some categories are merged. Finally, we presented a conceptual optimization model to incorporate the data-driven modeling approach for evacuation behavior into a government-led evacuation planning framework to improve the compliance rate. 

In this paper, we aim to address a relevant estimation problem that aviation professionals encounter in their daily operations. Specifically, aircraft load planners require information on the expected number of checked bags for a flight several hours prior to its scheduled departure to properly palletize and load the aircraft. However, the checked baggage prediction problem has not been sufficiently studied in the literature, particularly at the flight level. Existing prediction approaches have not properly accounted for the different impacts of overestimating and underestimating checked baggage volumes on airline operations. Therefore, we propose a custom loss function, in the form of a piecewise quadratic function, which aligns with airline operations practice and utilizes machine learning algorithms to optimize checked baggage predictions incorporating the new loss function. We consider multiple linear regression, LightGBM, and XGBoost, as supervised learning algorithms. We apply our proposed methods to baggage data from a major airline and additional data from various U.S. government agencies. We compare the performance of the three customized supervised learning algorithms. We find that the two gradient boosting methods (i.e., LightGBM and XGBoost) yield higher accuracy than the multiple linear regression; XGBoost outperforms LightGBM while LightGBM requires much less training time than XGBoost. We also investigate the performance of XGBoost on samples from different categories and provide insights for selecting an appropriate prediction algorithm to improve baggage prediction practices. Our modeling framework can be adapted to address other prediction challenges in aviation, such as predicting the number of standby passengers or no-shows.