A Stackelberg Framework for Resource-Aware LLM Agents: Learning, Repair, and Conditional Guarantees
On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks
Abstract (arXiv)
Large language model (LLM) agents increasingly operate as multi-turn systems that must allocate context, prompt verbosity, and tool access under finite computational budgets. Static thresholds are simple, but they are brittle under heterogeneous tasks and evolving session states. We formulate resource governance as a contextual Stackelberg game: a controller commits to a quality target and a cost incentive, while an executor responds with resource actions over context, prompting, and tool usage. We learn a conditional response model, optimize a leader policy against that model, and repair the resulting policy using real-API calibration and projection onto an empirically selected action set. For the restricted game, we establish conditional guarantees for equilibrium existence, follower-response stability, safe-set projection, and transfer from a surrogate environment to the real environment under bounded value error. The primary real-API experiment comprises 300 evaluated turns. Relative to a conservative baseline, the selected repaired controller reduces mean token cost by 17.4% (Welch p=0.022), while the measured quality difference is not statistically significant (p=0.44). The theoretical results are conditional and the experiments do not estimate their regret or transfer constants; consequently, the evidence establishes a promising repaired operating point, not a certified real-system equilibrium.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 16 of its 27 references that are in the map. First posted on arXiv 22 Jun 2026.
Builds on
16 of 27 references.
References · 16
- 2020Language Models are Few-Shot Learnerscited by 32,419
- 2017Proximal Policy Optimization Algorithmscited by 15,215
- 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
- 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
- 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
- 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
- 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
- 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,191
- 2016Generative Adversarial Imitation Learningcited by 1,597
- 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,498
- 2017Constrained Policy Optimizationcited by 983
- 2023FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performancecited by 853
4 more
- 2024WildChat: 1M ChatGPT Interaction Logs in the Wildcited by 629
- 2016Guided Cost Learning: Deep Inverse Optimal Control via Policy Optimizationcited by 406
- 2015Doubly Robust Off-policy Value Evaluation for Reinforcement Learningcited by 339
- 2017Effective Evaluation using Logged Bandit Feedback from Multiple Loggerscited by 20
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026From Task Solving to Robust Real-World Adaptation in LLM Agents
- 2026Agent libOS: A Runtime Substrate for Capability-Controlled Self-Evolving LLM Agents
- 2026Utility-Guided Agent Orchestration for Efficient LLM Tool Use
- 2026Scaling Agentic Capabilities, Not Context: Efficient Reinforcement Finetuning for Large Toolspaces
- 2026ClawArena-Team: Benchmarking Subagent Orchestration and Dynamic Workflows in Language-Model Agents
- 2026A Theory of LLM Information Susceptibility
- 2024Practical Considerations for Agentic LLM Systems
- 2025LLMs Working in Harmony: A Survey on the Technological Aspects of Building Effective LLM-Based Multi Agent Systems