Learning Drivers’ Utility Functions in a Coordinated Freight Routing System Based on Drivers’ Actions
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This study addresses the problem of traffic congestion in urban areas, specifically focusing on the inefficiencies caused by uncoordinated freight routing. The authors propose a coordinated freight routing system that utilizes personalized incentives to guide truck drivers toward routes that optimize overall network utility, moving beyond the suboptimal User Equilibrium (UE) where drivers act solely in self-interest. The motivation stems from the significant economic and environmental costs of congestion, particularly the disproportionate impact of trucks on traffic flow. The system aims to encourage voluntary participation by ensuring drivers receive higher expected utility, including incentives, than they would under UE conditions. The methodology employs a mixed logit model with a linear utility specification to capture driver route choice behaviors, accounting for attributes such as travel time, distance, and speed limits. To manage computational complexity, the system clusters drivers based on estimated utility parameters using a K-means algorithm. A central coordinator then solves a cluster-based optimization problem to assign routes and payments that minimize a weighted sum of passenger vehicle travel time and negative truck driver utility, subject to budget constraints. Personalized incentives are distributed using a greedy heuristic to maximize individual compliance. Crucially, the system dynamically updates utility parameter estimates using a hierarchical Bayes estimator based on historical driver choice data, allowing for continuous adaptation to driver preferences. Experimental validation was conducted using the Sioux Falls network. The results demonstrate that the coordinated routing system significantly improves network performance compared to UE scenarios. Sensitivity analyses reveal that system effectiveness is influenced by the volume of historical choice records, the number of Origin-Destination pairs, the number of driver clusters, and the available budget. The study finds that personalized incentives yield better compliance and system performance than cluster-based incentives alone. Furthermore, the system remains effective even with limited budget constraints, though additional funds further enhance performance. The dynamic learning component ensures that routing suggestions remain relevant as driver behaviors evolve. The significance of this work lies in its demonstration that personalized, data-driven incentive mechanisms can effectively mitigate congestion by aligning individual driver preferences with system-wide efficiency goals. By integrating behavioral modeling with real-time optimization and continuous learning, the proposed framework offers a scalable solution for freight management. The findings suggest that such coordinated systems can reduce travel times for passenger vehicles and improve overall network utility, providing a viable alternative to traditional congestion pricing or static routing methods. This approach highlights the potential of leveraging connected devices and big data to create adaptive transportation management systems that respect driver autonomy while achieving collective benefits.
Key finding
A coordinated freight routing system that dynamically learns driver utility functions and distributes personalized incentives significantly improves network efficiency and reduces congestion compared to user equilibrium conditions.
Methodology
modeling
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 24 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Theoretical Contribution: computational model