Paperlane

no. 16 · Tuesday 17 November 2026

A Fast and Accurate Time-Space Decomposition for Unit Commitment With Storage

Chance Brainard, Kylind Reagan and Shaked Regev

math.OC · 5 pp

Abstract

Grid-scale batteries couple decisions across time. This makes unit commitment (UC) with thousands of buses, batteries, generators, and alternating current constraints intractable. We introduce a time-space decomposition (TSD) heuristic to solve the problem. We first calculate target net battery charge or discharge rates on a zonal level, relaxing spatial constraints. We then incorporate this target into the Relax-and-Round UC (RRUC) solver from our previous work. RRUC enforces the spatial transmission constraints and transfers power between zones, if that reduces costs. We show that our TSD produces an error of about 0.1% on a simpler economic dispatch problem with batteries. This is larger than the error produced by RRUC in the tests in our previous work. So, we conjecture that the combined error of our two approximations is not much larger than 0.1%. We show that TSD with RRUC can solve systems with 10,000 buses in about 6 minutes, with sub-quadratic scaling, making it suitable for real time operations on an ISO-sized system.

comments
5 pages, 5 figures
fields
Optimization and Control
pdf ↗ page arXiv:2611.18946 ↗