arXiv: Van Ham, Liu and Wang propose battery dispatch framework

arXiv: Van Ham, Liu and Wang propose battery dispatch framework

Matthew Van Ham, Yiping Liu and Xiaozhe Wang propose a real-time framework for dispatching battery energy storage systems at data centers. It uses polynomial diffusion models to forecast electricity prices and data-center loads, then integrates those forecasts into dynamic programming to minimize electricity procurement costs. The authors describe batteries as a way to shift electricity purchases from high-price to low-price periods, against a backdrop of rising data-center electricity demand, market volatility and operating costs. Across the battery configurations they evaluated, they report 2.63 times the cost savings of a rule-based strategy. This is the paper’s reported comparison; the abstract does not give the evaluation conditions or the size of the savings in absolute terms. The supplied evidence describes a proposed forecasting and dispatch approach and its reported comparison, but does not show how the result would carry over to other settings.

Publication: 0580f771-b35f-48f3-8c29-0bbdfc95c3bd

attribution — pass: The reading attributes the framework and the 2.63x result to Matthew Van Ham, Yiping Liu and Xiaozhe Wang, matching the arXiv listing. The results are kept as the authors' reported comparison, not as an established fact.

Authors:Matthew Van Ham, Yiping Liu, Xiaozhe Wang

summary — uncertain: The summary is accurate but leaves the central term 'polynomial diffusion models' unexplained. AI-focused readers may confuse it with generative-AI diffusion models, though here it means a class of statistical models used for forecasting. 'Dispatch' (in the title), 'dynamic programming' and 'rule-based strategy' are also unexplained. The summary also omits the abstract's stated reason for using these models: their closed-form conditional expectations make the forecasts usable in the optimization. Smallest repair: add term-library tooltips for these terms, state the closed-form rationale, and plain-language the title, for example 'decide when data-center batteries charge and discharge'.

Leveraging the closed-form conditional expectations of PDMs, the forecasts are integrated into a dynamic programming (DP) framework

date — pass: The originalPublished date of 2026-11-05 matches the v1 submission date on arXiv. The feed date of Nov 6 is a later announcement.

[Submitted on 5 Nov 2026]

support — pass: Each claim matches the abstract verbatim. The 2.63x figure is kept as relative to a rule-based strategy across the evaluated configurations. The limitations accurately note the missing evaluation conditions, absolute savings and generalization. The question follows from those gaps without implying unsupported benefits.

Across the evaluated BESS configurations, the proposed approach achieves 2.63 times the cost savings of a rule-based (RB) strategy.

media — pass: The selected media is the verified arXiv abstract page for this paper. It is the direct original source, and no better media was found.