Shippers’ Behavior Study Through Developing and Calibrating Their Utility Functions

Details

• Creators:
  Han, Lee ; Wang, Bruce ; Zhang, Yunlong ; Mohammadian, Kouros
• Corporate Creators:
  Center for Freight Transportation for Efficient and Resilient Supply Chain (FERSC) Tier 1 University Transportation Center (UTC)
• Corporate Contributors:
  United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Federal Highway Administration ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology
• Subject/TRT Terms:
  Analysis Behavior Calibration Commodity Flow Decision Making Freight Handling Logits Machine Learning Mobility Multinomial Logits Shipping Surveying Transportation Modes Transportation Planning

  More +
• Publication/ Report Number:
  FERSC-2023-Project5-1
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  9/1/2023–8/15/2024
• Corporate Publisher:
  Center for Freight Transportation for Efficient & Resilient Supply Chain
• Abstract:
  This study addresses the lack of granularity in traditional freight demand models by developing a comprehensive survey to collect disaggregated data on freight movements and the underlying logistics decisions made by shippers. The primary objective was to enhance the precision of freight demand forecasts and gain a deeper understanding of the complexities within freight transportation networks. The survey targeted logistics managers and shippers, gathering detailed information about their firms, including the types and volumes of goods, as well as the specific characteristics of individual shipments. This disaggregated data enabled the construction of utility-based models that examine shipper decision-making from a granular perspective, revealing the factors influencing transportation mode and route selection. Building upon the survey insights, the study implemented several key advancements: (1) developing localized models for individual commodity or industry categories; (2) incorporating geographical features, such as distance between origin and destination zones, to improve predictive accuracy; (3) comparing the performance of machine learning ensemble models and utility function-based approaches, including multinomial logistic regression and nonlinear classification models, achieving over 92% accuracy in mode prediction; and (4) creating a freight transportation dashboard to visualize and interpret key transportation data and trends. Addressing the common issue of data imbalance in freight logistics, the study employed machine learning techniques to manage skewed distributions across different commodities and transportation choices. These efforts collectively enhance the ability to predict and interpret freight movement, supporting the planning and operations of more efficient multimodal transportation systems. The findings from this study provide a foundation for improving freight demand forecasting and transportation planning, enabling stakeholders to make more informed decisions that promote sustainable and effective freight mobility across the nation.
• Format:
  PDF
• Funding:
  69A3552348338
• Collection(s):
  University Transportation Centers
• Main Document Checksum:
  urn:sha-512:bdbe53f368f2dcc094b4b840258eb7fdc0cf5c9b70bb0a8a75ffb5257339ac327badde179be90ecafbb143682889564d1fde122f5011739c0a1102831fcf2a9c
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/86964/dot_86964_DS1.pdf
• File Type:
  [PDF - 2.24 MB ]