Co-Saving: Resource Aware Multi-Agent Collaboration for Software Development

2025-05-28

AI Summary

Current multi-agent systems (MAS) for complex software development tasks lack resource awareness, resulting in excessive token consumption and high response latency. To address this, we propose a resource-aware multi-agent collaboration framework centered on a novel “shortcut” mechanism: by modeling experiential knowledge from historical successful trajectories, the framework dynamically bypasses redundant reasoning steps and agent interactions while preserving solution quality. Our approach integrates an LLM-driven multi-agent architecture, role-based task decomposition, trajectory-guided experience modeling, and adaptive communication scheduling. Experimental results demonstrate that, compared to ChatDev, our framework reduces average token usage by 50.85% and improves functional correctness of generated code by 10.06%. These gains significantly enhance the practicality and scalability of MAS in resource-constrained environments.

Technology Category

• Multiagent Systems: Agent Communication
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Recent advancements in Large Language Models (LLMs) and autonomous agents have demonstrated remarkable capabilities across various domains. However, standalone agents frequently encounter limitations when handling complex tasks that demand extensive interactions and substantial computational resources. Although Multi-Agent Systems (MAS) alleviate some of these limitations through collaborative mechanisms like task decomposition, iterative communication, and role specialization, they typically remain resource-unaware, incurring significant inefficiencies due to high token consumption and excessive execution time. To address these limitations, we propose a resource-aware multi-agent system -- Co-Saving (meaning that multiple agents collaboratively engage in resource-saving activities), which leverages experiential knowledge to enhance operational efficiency and solution quality. Our key innovation is the introduction of"shortcuts"-- instructional transitions learned from historically successful trajectories -- which allows to bypass redundant reasoning agents and expedite the collective problem-solving process. Experiments for software development tasks demonstrate significant advantages over existing methods. Specifically, compared to the state-of-the-art MAS ChatDev, our method achieves an average reduction of 50.85% in token usage, and improves the overall code quality by 10.06%.

Problem

• Addresses high token consumption in multi-agent systems
• Reduces redundant reasoning in collaborative problem-solving
• Improves efficiency and code quality in software development

Innovation

• Resource-aware multi-agent collaboration system
• Shortcuts from successful historical trajectories
• Reduces token usage and improves code quality