Global-Decision-Focused Neural ODEs for Proactive Grid Resilience Management

TL;DR

This paper introduces a novel neural ODE framework called GDF that integrates outage prediction with global optimization for proactive power grid resilience, improving decision coherence and operational efficiency.

Contribution

The paper presents the GDF neural ODE model that jointly predicts outages and optimizes resilience strategies in a decision-aware, globally coherent manner.

Findings

Enhanced outage prediction consistency demonstrated on real datasets

Improved grid resilience through integrated prediction and optimization

Significant operational efficiency gains over traditional two-stage methods

Abstract

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.