Learning Latency-Aware Orchestration for Parallel Multi-Agent Systems

TL;DR

This paper introduces LAMaS, a learning-based framework that optimizes latency in parallel multi-agent systems, significantly reducing critical path length while maintaining or improving task performance.

Contribution

It presents a novel latency-aware orchestration framework for parallel multi-agent systems that explicitly optimizes execution latency and constructs efficient execution topologies.

Findings

Reduces critical path length by 38-46% compared to baselines.

Maintains or improves task performance with latency optimization.

Demonstrates effectiveness across multiple benchmarks.

Abstract

Multi-agent systems (MAS) enable complex reasoning by coordinating multiple agents, but often incur high inference latency due to multi-step execution and repeated model invocations, severely limiting their scalability and usability in time-sensitive scenarios. Most existing approaches primarily optimize task performance and inference cost, and explicitly or implicitly assume sequential execution, making them less optimal for controlling latency under parallel execution. In this work, we investigate learning-based orchestration of multi-agent systems with explicit latency supervision under parallel execution. We propose Latency-Aware Multi-agent System (LAMaS), a latency-aware multi-agent orchestration framework that enables parallel execution and explicitly optimizes the critical execution path, allowing the controller to construct execution topology graphs with lower latency under parallel execution. Our experiments show that our approach reduces critical path length by 38-46% compared to the state-of-the-art baseline for multi-agent architecture search across multiple benchmarks, while maintaining or even improving task performance. These results highlight the importance of explicitly optimizing latency under parallel execution when designing efficient multi-agent systems. The code is available at https://github.com/xishi404/LAMaS