Self-Resource Allocation in Multi-Agent LLM Systems

2025-04-02

AI Summary

This study addresses computational task allocation in multi-agent LLM systems, balancing cost, efficiency, and performance. Methodologically, it integrates an LLM-based meta-reasoning framework, hierarchical task decomposition, role-aware prompting, and concurrent multi-agent action modeling. Its key contributions are threefold: (1) the first systematic comparative analysis of LLMs acting as *planners* versus *orchestrators* in resource scheduling; (2) the novel explicit modeling of worker capabilities to enhance allocation robustness; and (3) empirical validation showing the planner paradigm significantly outperforms the orchestrator—improving both task throughput and agent utilization under concurrency. Experiments further demonstrate that explicit capability prompting boosts task assignment accuracy by 23.6% in suboptimal-worker scenarios. Collectively, the work advances principled, scalable, and robust task orchestration in LLM-based multi-agent systems.

Technology Category

• Multiagent Systems: Multiagent Planning
• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

With the development of LLMs as agents, there is a growing interest in connecting multiple agents into multi-agent systems to solve tasks concurrently, focusing on their role in task assignment and coordination. This paper explores how LLMs can effectively allocate computational tasks among multiple agents, considering factors such as cost, efficiency, and performance. In this work, we address key questions, including the effectiveness of LLMs as orchestrators and planners, comparing their effectiveness in task assignment and coordination. Our experiments demonstrate that LLMs can achieve high validity and accuracy in resource allocation tasks. We find that the planner method outperforms the orchestrator method in handling concurrent actions, resulting in improved efficiency and better utilization of agents. Additionally, we show that providing explicit information about worker capabilities enhances the allocation strategies of planners, particularly when dealing with suboptimal workers.

Problem

• Allocate computational tasks among multiple LLM agents effectively
• Compare LLM effectiveness as orchestrators and planners
• Enhance allocation strategies with explicit worker capability information

Innovation

• LLMs allocate tasks among multiple agents effectively
• Planner method excels in concurrent action handling
• Explicit worker capability info improves allocation strategies