Cherry-Picking Load Shaping for More Effective Carbon Reduction

Researchers led by Bokan Chen have submitted an arXiv paper proposing a ‘cherry-picking’ daily load-shaping approach to reduce grid CO2 emissions and electricity cost (arXiv:2601.17990, submitted 25 Jan 2026).

• Main announcement: The paper proposes a cherry-picking strategy that selects a daily load-shaping policy based on observable grid signals and historical data, and demonstrates via calibrated ERCOT day-ahead DC-OPF simulations that this approach can materially improve on common strategies (including LMP-based shaping). The study focuses on multi-megawatt loads such as data centers and other large flexible consumers (VPPs, distributed energy resources).
• Background & details: The submission is v1 on arXiv (25 Jan 2026); full PDF and TeX source are available on arXiv, the paper is categorized under Machine Learning (stat.ML, cs.LG) and Systems and Control (eess.SY), and the work is released under a Creative Commons BY-SA 4.0 license; a DataCite/arXiv-issued DOI link is provided (pending registration).