Uncertainty-Aware Resource Orchestration for Agentic AI Systems

Computer Science > Artificial Intelligence arXiv:2611.05392 (cs) [Submitted on 14 Sep 2026] Title: Uncertainty-Aware Resource Orchestration for Agentic AI Systems Authors: Masoud Shokrnezhad, Tarik Taleb.

计算机科学 > 人工智能 arXiv:2611.05392 (cs) [提交于 2026 年 9 月 14 日] 标题：面向智能体 AI 系统的资源编排（具备不确定性感知能力） 作者：Masoud Shokrnezhad, Tarik Taleb。

Abstract: Agentic artificial intelligence (AI) systems powered by large language models (LLMs) execute as autonomous multi-step pipelines that dynamically invoke tools, integrate observations into memory, and iterate until a task is resolved or a resource budget is exhausted. Unlike classical workflows with fixed graphs, the aggregate compute, bandwidth, and end-to-end latency demands of such pipelines are unknown at dispatch time: they emerge at runtime from intermediate observations and stochastic token sampling, yielding heavy-tailed, multi-modal random variables.

摘要：由大语言模型（LLM）驱动的智能体人工智能（AI）系统以自主的多步流水线形式执行，动态调用工具、将观察结果整合到记忆中，并进行迭代，直到任务解决或资源预算耗尽。与具有固定图结构的传统工作流不同，此类流水线在调度时的总计算量、带宽和端到端延迟需求是未知的：它们在运行时根据中间观察结果和随机标记采样产生，从而形成具有重尾分布和多模态特征的随机变量。

This unpredictability forces infrastructure orchestrators into a dilemma: under-provisioning risks mid-execution throttling and deadline violations, while over-provisioning wastes scarce edge or cloud capacity. Existing schedulers require a fully known pipeline graph, while edge resource-management methods rely on parametric or stationary demand models, and neither fits the unknown, non-stationary profiles of agentic pipelines.

这种不可预测性使基础设施编排器陷入了两难境地：供应不足会导致执行中途受限和违反截止日期，而过度供应则会浪费稀缺的边缘或云端容量。现有的调度程序需要完全已知的流水线图，而边缘资源管理方法依赖于参数化或平稳的需求模型，两者都不适用于智能体流水线未知且非平稳的特征。

We formalize this setting as the uncertainty-aware agentic orchestration problem, where the goal is to reserve, at dispatch time, compute and bandwidth allocations that cover actual pipeline demands, and to verify that the pipeline can meet its deadline, each with a prescribed probability. We then propose Conformal Reservation for Orchestration (CRO), an online algorithm that translates a sliding window of past execution traces into dispatch-time resource allocations with finite-sample, distribution-free coverage guarantees via conformal prediction.

我们将此场景形式化为“不确定性感知智能体编排问题”，其目标是在调度时预留能够覆盖实际流水线需求的计算和带宽分配，并以预设的概率验证流水线能否满足截止日期。随后，我们提出了“编排共形预留”（CRO），这是一种在线算法，通过共形预测将过去执行轨迹的滑动窗口转化为具有有限样本、无分布覆盖保证的调度时资源分配。

Simulations over heterogeneous task mixes with multi-modal and heavy-tailed profiles show that CRO is the only evaluated method that simultaneously satisfies target probabilistic coverage across all risk levels while, among coverage-feasible methods, attaining the lowest over-reservation, roughly halving the resource waste of the Chebyshev-bound baseline.

针对具有多模态和重尾特征的异构任务组合进行的模拟表明，CRO 是唯一一种在所有风险水平下均能同时满足目标概率覆盖率的评估方法；在所有可实现覆盖的方法中，它实现了最低的过度预留，将基于切比雪夫界限的基准方法的资源浪费减少了约一半。