BAMAS: Structuring Budget-Aware Multi-Agent Systems

2025-11-26

AI Summary

Large language model (LLM)-driven multi-agent systems incur prohibitively high inference costs under budget constraints. Method: This paper proposes the first end-to-end budget-aware joint optimization framework, unifying LLM selection and multi-agent collaboration topology design. It employs integer linear programming for cost-effective, high-performance LLM candidate screening and integrates reinforcement learning to dynamically optimize interaction structures and instantiation strategies. Contribution/Results: Unlike prior approaches that decouple model selection from collaboration design, our framework jointly optimizes both components. Experiments across three representative task domains demonstrate that the method achieves performance on par with state-of-the-art (SOTA) baselines while reducing inference cost by up to 86%. This significantly improves budget utilization efficiency and practical deployment feasibility.

Technology Category

Multiagent Systems: Multiagent LearningMachine Learning: Large Multimodal Models (LMMs)Search and Optimization: Learning to Search

Application Category

Economics, Online Markets and Human Computation: Cost models of using LLMs in production systemsUser Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendationSemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large language model (LLM)-based multi-agent systems have emerged as a powerful paradigm for enabling autonomous agents to solve complex tasks. As these systems scale in complexity, cost becomes an important consideration for practical deployment. However, existing work rarely addresses how to structure multi-agent systems under explicit budget constraints. In this paper, we propose BAMAS, a novel approach for building multi-agent systems with budget awareness. BAMAS first selects an optimal set of LLMs by formulating and solving an Integer Linear Programming problem that balances performance and cost. It then determines how these LLMs should collaborate by leveraging a reinforcement learning-based method to select the interaction topology. Finally, the system is instantiated and executed based on the selected agents and their collaboration topology. We evaluate BAMAS on three representative tasks and compare it with state-of-the-art agent construction methods. Results show that BAMAS achieves comparable performance while reducing cost by up to 86%.

Problem

• Structuring multi-agent systems under explicit budget constraints
• Selecting optimal LLM sets balancing performance and cost
• Determining collaboration topology through reinforcement learning methods

Innovation

• Selects optimal LLMs via Integer Linear Programming
• Uses reinforcement learning for interaction topology
• Instantiates system with chosen agents and topology