Locked In：面向 AI 数据中心建设的多目标强化学习方法
Locked In: Multi-Objective RL for the AI Data Center Build-Out
AI 导读
研究者将 AI 数据中心选址与建设这一数万亿美元、长达 10 到 20 年资产周期的长期规划形式化为分层多目标 Markov 决策过程，用偏好条件化的多目标智能体（semi-Markov manager-worker 层级）求解，使规划者能在推理时跨成本、碳、水与可靠性调整策略。
正文 · 原文
Abstract:AI data center build-out is a multi-trillion-dollar, multi-decade infrastructure program. Yet siting, power procurement, and cooling decisions are still made with static, single-point scoring tools. These tools are evaluated once, on day-zero numbers. This is a poor fit. A site's water, power, and carbon profile drifts substantially over the 10 to 20 year asset life. The governing decisions are also irreversible, and they bind capital and physical plant for years. We formalize long-horizon build-out as a tiered multi-objective Markov decision process. Strategic commitments are gated to discrete windows, incur switching costs, and energize only after source-dependent lead times. Operational dispatch is restricted to capacity that physically exists. We solve the problem with a preference-conditioned multi-objective agent that has a semi-Markov manager and worker hierarchy. The agent recovers a family of policies that a planner can steer across cost, carbon, water, and reliability at inference time. We validate long-horizon dynamics with a lightweight counterfactual rollout model. Across three commitment regimes (irreversible, reversible, and commit-once) on real site data, preference conditioning yields a confirmed, steerable trade-off on both objective pairs in every regime. Imposing lock-in produces a more evenly covered frontier and a far more consistent rate of anticipatory retrofit ahead of a binding water-stress constraint than a freely reversible formulation. Finally, a compact counterfactual dynamics model tracks slow-moving state well. It is measurably outperformed by simple persistence on the water-stress axis. We report this limitation rather than obscure it. We release the environment, agents, baselines, and evaluation harness. All of them are grounded in real interconnection-queue, water-stress, and grid-emissions data.
| Comments: | Tackling Climate Change with Machine Learning Workshop at NeurIPS 2026 |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.35147 [cs.LG] |
| (or arXiv:2610.35147v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.35147 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Satyapragnya Kar [view email]
[v1]
Mon, 26 Oct 2026 10:14:37 UTC (3,218 KB)
来源：arXiv cs.AI · arxiv.org