SCALFLEX: A Scalable Heuristic Framework for Distributed Energy Flexibility and Grid Tariff Optimization

Paper in proceeding, 2026

The increasing deployment of distributed energy resources is rapidly expanding the flexibility available in power systems (e.g., shifting consumption or storage in response to price signals), creating significant computational challenges for large-scale coordination. We present ScalFlex, a scalable heuristic framework for efficient optimization of distributed flexibility. ScalFlex approximates new optimization tasks by reusing previously computed solutions for similar load profiles, enabling high-throughput scheduling with limited optimality loss. We evaluate the framework in two applications: (i) large-scale household battery scheduling under dynamic pricing, and (ii) the computation of hourly grid tariffs to reduce peak demand through coordinated flexibility shifts. Using data from over 2,000 households and more than 100,000 optimization tasks, ScalFlex achieves up to 18 × higher throughput than a linear programming baseline while remaining within a few percent of optimality. These results show that computation-aware heuristics enable scalable coordination of distributed flexibility and tariff-driven mechanisms.