Adaptive routing algorithms for transportation networks using Partially Observable Markov Decision Processes
Student Classification
Anusha Neupane, 4th year, Civil Engineering
Faculty Mentor
Dr. Venktesh Pandey, Department of Civil, Architectural and Environmental Engineering
Department
Department of Civil, Architectural and Environmental Engineering
Document Type
Poster
Publication Date
Spring 2023
Abstract
Express lane facilities provide toll and travel time information in real- time through variable message signs or phone apps. In contrast to the assumptions in the adaptive routing literature, the online travel time or toll information is commonly incomplete, error-prone, or aggregated for the entire route. In this research, we investigate algorithms that find the least-expected-cost routing strategy from an origin to a destination given partial information at different diverge locations. My undergraduate research has focused on generate the test datasets including tasks such as (a) analysis of travel time inventory using National Performance Management Research DataSet (NPMRDS), (b) programming the network to integrate mean and standard deviation of travel time attributes, and (c) testing the partially observable Markov decision process (POMDP)-based iterative algorithms to solve the adaptive routing problems. The suggested routing algorithms show promising results and can be used in express lane planning and pricing models as well as adaptive route-recommender software.
Recommended Citation
Neupane, Anusha, "Adaptive routing algorithms for transportation networks using Partially Observable Markov Decision Processes" (2023). Undergraduate Research and Creative Inquiry Symposia. 308.
https://digital.library.ncat.edu/ugresearchsymposia/308