# Learning to Assign Prediction Tasks to Agents with Capacity Constraints

arXiv: https://arxiv.org/abs/2605.27999 · HTML: https://mlsanity.com/paper/2605.27999/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Agent skill learning and evolution, one of 16 directions in LLM agents and tool use.
- Builds on 26 of its 51 references that are in the map.
- First posted on arXiv 27 May 2026.

## Placed in

- LLM agents and tool use / [Agent skill learning and evolution](https://mlsanity.com/agents/agent-skill-learning-and-evolution/)

## Builds on (most cited first, 26 of 26)

- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance](https://mlsanity.com/paper/2305.05176/) (2023)
- [RouteLLM: Learning to Route LLMs with Preference Data](https://mlsanity.com/paper/2406.18665/) (2024)
- [Consistent Estimators for Learning to Defer to an Expert](https://mlsanity.com/paper/2006.01862/) (2020)
- [Predict Responsibly: Improving Fairness and Accuracy by Learning to Defer](https://mlsanity.com/paper/1711.06664/) (2017)
- [Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy](https://mlsanity.com/paper/2002.04087/) (2020)
- [Bandits with Knapsacks](https://mlsanity.com/paper/1305.2545/) (2013)
- [Towards Unbiased and Accurate Deferral to Multiple Experts](https://mlsanity.com/paper/2102.13004/) (2021)
- [Who Should Predict? Exact Algorithms For Learning to Defer to Humans](https://mlsanity.com/paper/2301.06197/) (2023)
- [Combinatorial Sleeping Bandits with Fairness Constraints](https://mlsanity.com/paper/1901.04891/) (2019)
- [Human-AI Collaboration via Conditional Delegation: A Case Study of Content Moderation](https://mlsanity.com/paper/2204.11788/) (2022)
- [Dual Mirror Descent for Online Allocation Problems](https://mlsanity.com/paper/2002.10421/) (2020)
- [Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing](https://mlsanity.com/paper/2405.00467/) (2024)
- [Fair Contextual Multi-Armed Bandits: Theory and Experiments](https://mlsanity.com/paper/1912.08055/) (2019)
- [IRT-Router: Effective and Interpretable Multi-LLM Routing via Item Response Theory](https://mlsanity.com/paper/2506.01048/) (2025)
- [The Unreasonable Effectiveness of Greedy Algorithms in Multi-Armed Bandit with Many Arms](https://mlsanity.com/paper/2002.10121/) (2020)
- [Adaptive LLM Routing under Budget Constraints](https://mlsanity.com/paper/2508.21141/) (2025)
- [Thompson sampling with the online bootstrap](https://mlsanity.com/paper/1410.4009/) (2014)
- [Lifting the Information Ratio: An Information-Theoretic Analysis of Thompson Sampling for Contextual Bandits](https://mlsanity.com/paper/2205.13924/) (2022)
- [Budgeted Multiple-Expert Deferral](https://mlsanity.com/paper/2510.26706/) (2025)
- [Cost-Sensitive Learning to Defer to Multiple Experts with Workload Constraints](https://mlsanity.com/paper/2403.06906/) (2024)
- [Delay-Adaptive Learning in Generalized Linear Contextual Bandits](https://mlsanity.com/paper/2003.05174/) (2020)
- [Online Decision Deferral under Budget Constraints](https://mlsanity.com/paper/2409.20489/) (2024)
- [ParetoBandit: Budget-Paced Adaptive Routing for Non-Stationary LLM Serving](https://mlsanity.com/paper/2604.00136/) (2026)
- [Thompson Sampling for Combinatorial Semi-bandits with Sleeping Arms and Long-Term Fairness Constraints](https://mlsanity.com/paper/2005.06725/) (2020)
- [Fatigue-Aware Learning to Defer via Constrained Optimisation](https://mlsanity.com/paper/2604.00904/) (2026)

## Abstract (arXiv)

We address the problem of learning to assign prediction tasks to one agent from a set of available agents, including human decision-makers and AI models. We focus on sequential learning of agent expertise and assignment policies where each agent is constrained to handle a fraction of tasks. We provide a general theoretical characterization of this problem in terms of agent capacities, differences in agent expertise, and task context. We then develop a framework of sequential explore-exploit policy-learning algorithms that seek to maximize overall performance. Experimental results over a variety of tabular, image, and text prediction tasks demonstrate systematic gains from our policy-learning algorithms relative to non-contextual baselines across different types of agents, including LLMs and humans.