English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present a multi-agent Deep Reinforcement Learning (DRL) framework formanaging large transportation infrastructure systems over their life-cycle.Life-cycle management of such engineering systems is a computationallyintensive task, requiring appropriate sequential inspection and maintenancedecisions able to reduce long-term risks and costs, while dealing withdifferent uncertainties and constraints that lie in high-dimensional spaces. Todate, static age- or condition-based maintenance methods and risk-based orperiodic inspection plans have mostly addressed this class of optimizationproblems. However, optimality, scalability, and uncertainty limitations areoften manifested under such approaches. The optimization problem in this workis cast in the framework of constrained Partially Observable Markov DecisionProcesses (POMDPs), which provides a comprehensive mathematical basis forstochastic sequential decision settings with observation uncertainties, riskconsiderations, and limited resources. To address significantly large state andaction spaces, a Deep Decentralized Multi-agent Actor-Critic (DDMAC) DRL methodwith Centralized Training and Decentralized Execution (CTDE), termed asDDMAC-CTDE is developed. The performance strengths of the DDMAC-CTDE method aredemonstrated in a generally representative and realistic example application ofan existing transportation network in Virginia, USA. The network includesseveral bridge and pavement components with nonstationary degradation,agency-imposed constraints, and traffic delay and risk considerations. Comparedto traditional management policies for transportation networks, the proposedDDMAC-CTDE method vastly outperforms its counterparts. Overall, the proposedalgorithmic framework provides near optimal solutions for transportationinfrastructure management under real-world constraints and complexities.

Multi-agent deep reinforcement learning with centralized training and decentralized execution for transportation infrastructure management