Global-Decision-Focused Neural ODEs for Proactive Grid Resilience Management
This addresses the problem of suboptimal resource allocation in power systems during disasters, offering a domain-specific incremental advance.
The paper tackles the misalignment between prediction and optimization in grid operations for extreme hazard events by proposing a predict-all-then-optimize-globally (PATOG) framework, resulting in significant improvements in outage prediction consistency and grid resilience.
Extreme hazard events such as wildfires and hurricanes increasingly threaten power systems, causing widespread outages and disrupting critical services. Recently, predict-then-optimize approaches have gained traction in grid operations, where system functionality forecasts are first generated and then used as inputs for downstream decision-making. However, this two-stage method often results in a misalignment between prediction and optimization objectives, leading to suboptimal resource allocation. To address this, we propose predict-all-then-optimize-globally (PATOG), a framework that integrates outage prediction with globally optimized interventions. At its core, our global-decision-focused (GDF) neural ODE model captures outage dynamics while optimizing resilience strategies in a decision-aware manner. Unlike conventional methods, our approach ensures spatially and temporally coherent decision-making, improving both predictive accuracy and operational efficiency. Experiments on synthetic and real-world datasets demonstrate significant improvements in outage prediction consistency and grid resilience.