SYMPHONY: Synergistic Multi-agent Planning with Heterogeneous Language Model Assembly

TL;DR

SYMPHONY introduces a multi-agent framework using diverse language models to improve exploration and performance in complex planning tasks, outperforming existing methods with open-source and cloud-based models.

Contribution

The paper presents a novel multi-agent planning approach that combines heterogeneous language models to enhance exploration and planning efficiency.

Findings

SYMPHONY outperforms state-of-the-art baselines on benchmark tasks.

Heterogeneous multi-agent coordination improves planning diversity.

Open-source LLMs are sufficient for strong performance, further enhanced by cloud-based models.

Abstract

Recent advancements have increasingly focused on leveraging large language models (LLMs) to construct autonomous agents for complex problem-solving tasks. However, existing approaches predominantly employ a single-agent framework to generate search branches and estimate rewards during Monte Carlo Tree Search (MCTS) planning. This single-agent paradigm inherently limits exploration capabilities, often resulting in insufficient diversity among generated branches and suboptimal planning performance. To overcome these limitations, we propose Synergistic Multi-agent Planning with Heterogeneous langauge model assembly (SYMPHONY), a novel multi-agent planning framework that integrates a pool of heterogeneous language model-based agents. By leveraging diverse reasoning patterns across agents, SYMPHONY enhances rollout diversity and facilitates more effective exploration. Empirical results across multiple benchmark tasks show that SYMPHONY achieves strong performance even when instantiated with open-source LLMs deployable on consumer-grade hardware. When enhanced with cloud-based LLMs accessible via API, SYMPHONY demonstrates further improvements, outperforming existing state-of-the-art baselines and underscoring the effectiveness of heterogeneous multi-agent coordination in planning tasks.