Advance the boundaries of post-earthquake transportation network restoration decision-making through advanced AI techniques, marking a major departure from existing studies that largely relied on…

The University of Texas at San Antonio (UTSA) was awarded a $399,922 Cooperative Agreement under the National Institute of Standards and Technology's (NIST) Measurement and Engineering Research and Standards (CFDA 11.609) program. The purpose of this 3-year award is to advance the use of deep reinforcement learning techniques for post-earthquake transportation network restoration decision-making. Key activities include: (1) developing a novel deep reinforcement learning agent for adaptive transportation network restoration; (2) creating a hierarchical deep reinforcement learning agent for large-scale transportation networks; and (3) integrating the deep reinforcement learning techniques with an interactive dashboard to enhance the transparency and confidence in the resulting AI policies. This research aims to facilitate more rapid and cost-effective post-earthquake transportation network restoration, thereby reducing the impact on regional recovery efforts. The project will be evaluated on a testbed in Shelby County, Tennessee. UTSA is the sole recipient under this Cooperative Agreement and does not intend to subaward any funds.

Purpose

The purpose of this grant is to advance the boundaries of post-earthquake transportation network restoration decision-making through advanced AI techniques, marking a major departure from existing studies that largely relied on ranking-based prioritization or deterministic optimizations.

Activities

This research will (1) develop a novel deep reinforcement learning agent for adaptive transportation network restoration decision-making; (2) develop a hierarchical deep reinforcement learning agent for complex large-scale transportation networks; and (3) integrate deep reinforcement learning techniques with an interactive dashboard for enhanced confidence and transparency of the AI policies. The proposed research will be evaluated on a regional testbed in shelby county, tennessee.

Expected outcomes

This research will facilitate more rapid and cost-effective post-earthquake TN restoration, thereby greatly reducing the interruption to the recovery of the regional built environment, business operation, and people's daily lives.

Intended beneficiaries

Local residents, businesses, and communities subject to transportation network damage by earthquake events.

Subrecipient activities

The recipient does not intend to subaward funds.