SwarmAgentic
Towards Fully Automated Agentic System Generation via Swarm Intelligence
SwarmAgentic is a framework for fully automated agentic system generation that constructs agentic systems from scratch and jointly optimizes agent functionality and collaboration as interdependent components through language-driven exploration.
Key Features
🛠️ Autonomous, from scratch
Builds complete multi-agent systems directly from task descriptions and an objective function, jointly optimizing both functionality and collaboration without predefined templates or seed workflows.
🐝 Language-driven PSO
Reformulates particle swarm optimization into interpretable text–symbol updates over agent roles and coordination structures, enabling language-mediated exploration of the system design space.
🔁 Failure-aware refinement
Leverages LLM-guided flaw detection and failure memory to prevent repeated suboptimal adjustments and focus updates on meaningful improvements.
🚀 Robust generalization
Demonstrates strong performance and transferability across diverse real-world, open-ended tasks requiring high-level planning, multi-agent coordination, and complex reasoning.
Results
SwarmAgentic outperforms all baselines, achieving a +261.8% relative improvement over ADAS on the TravelPlanner benchmark, highlighting the effectiveness of full automation in structurally unconstrained tasks.
Links
- Project Page: https://yaoz720.github.io/SwarmAgentic/
- arXiv: 2506.15672
- Code: GitHub Repository