Abstract:In the context of modern engineering, environmental, and societal concerns, there is an increasing demand for methods able to identify rational management strategies for civil engineering systems, minimizing structural failure risks while optimally planning inspection and maintenance (I&M) processes. Most available methods simplify the I&M decision problem to the component level, often assuming statistical, structural, or cost independence among components, due to the computational complexity associated with global optimization methodologies under joint system-level state descriptions. In this paper, we propose an efficient algorithmic framework for inference and decision-making under uncertainty for engineering systems exposed to deteriorating environments, providing optimal management strategies directly at the system level. In our approach, the decision problem is formulated as a factored partially observable Markov decision process, whose dynamics are encoded in Bayesian network conditional structures. The methodology can handle environments under equal or general, unequal deterioration correlations among components, through Gaussian hierarchical structures and dynamic Bayesian networks, decoupling the originally joint system state space to component networks conditional on shared random variables. In terms of policy optimization, we adopt a deep decentralized multi-agent actor-critic (DDMAC) reinforcement learning approach, in which the policies are approximated by actor neural networks guided by a critic network. By including deterioration dependence in the simulated environment, and by formulating the cost model at the system level, DDMAC policies intrinsically consider the underlying system-effects. This is demonstrated through numerical experiments conducted for both a 9-out-of-10 system and a steel frame under fatigue deterioration. Results demonstrate that DDMAC policies offer substantial benefits when compared to state-of-the-art heuristic approaches. The inherent consideration of system-effects by DDMAC strategies is also interpreted based on the learned policies.

Maintenance in Probabilistic Knowledge-Based Systems

Conditional Evidence Theory and Its Application in Knowledge Discovery

Incorporating Probabilistic Knowledge into Topic Models.

Computing and Maintaining Provenance of Query Result Probabilities in Uncertain Knowledge Graphs

An algebra of bayesian belief universes for knowledge-based systems

Inference and dynamic decision-making for deteriorating systems with probabilistic dependencies through Bayesian networks and deep reinforcement learning

The impact of prior knowledge on causal structure learning

Reliability of a probabilistic knowledge structure

Effective refractive index tensor for weak field gravity

Epitaph or Breaking News? Analyzing and Predicting the Stability of Knowledge Base Properties

Bayesian networks and probabilistic reasoning about scientific evidence when there is a lack of data

Errors are Robustly Tamed in Cumulative Knowledge Processes

Implementing Probabilistic Reasoning

Dynamic Network Updating Techniques For Diagnostic Reasoning

Correctness in hierarchical knowledge-based requirements

Qualitative Propagation and Scenario-based Explanation of Probabilistic Reasoning

A maintenance policy for a deteriorating system under epistemic uncertainty

Probabilistic confidence for decisions based on uncertain reliability estimates

Querying and Repairing Inconsistent Prioritized Knowledge Bases: Complexity Analysis and Links with Abstract Argumentation

Robust Bayesian Reliability for Complex Systems under Prior-Data Conflict

Uncertain Knowledge Representation And Inference For Tracing Errors In Uncertain Data