August 1, 2019.
The modeling of delay diffusion in airport networks can potentially help develop strategies to prevent the spread of such delays and disruptions. With this goal, we used the publicly-available historical United States Federal Aviation Administration (FAA) flight data to model the spread of delays in the US airport network. For the major (ASPM-77) airports for January 2017, using a threshold on the volume of flights, we sparsify the network in order to better recognize patterns and cluster structure of the network. We developed a diffusion simulator and greedy optimizer to find the top influential airport nodes that propagate the most delays in the entire network and in each cluster. Currently, our model is based on a popular diffusion model, Independent Cascade Model. We visualize the delay spread in order to better represent how the affected airports can spread these delays to the entire network. In the process, we also study inter-cluster propagation and intra-cluster propagation.
Databases and Information Systems | Numerical Analysis and Computation | Numerical Analysis and Scientific Computing | Other Applied Mathematics
Pacific Northwest National Laboratory (PNNL)
The 2019 STEM Teacher and Researcher Program and this project have been made possible through support from Chevron (http://www.chevron.com)/, the National Science Foundation through the Robert Noyce Program under Grant #1836335 and 1340110, the California State University Office of the Chancellor, and California Polytechnic State University in partnership with Pacific Northwest National Laboratory and Arun Sathanur.