EvolveRouter: Co-Evolving Routing and Prompt for Multi-Agent Question Answering

2026-04-06

AI Summary

Existing routing methods in multi-agent question answering struggle to jointly optimize agent capabilities and collaboration structures, often relying on fixed collaboration scales that cannot adapt to dynamic query demands. This work proposes EvolveRouter, a novel framework that introduces, for the first time, a closed-loop co-evolution mechanism combining graph neural network-based routing with instruction fine-tuning, alongside an adaptive reasoning strategy grounded in answer consistency. This approach simultaneously optimizes both agent competencies and collaboration topology while dynamically adjusting the number of participating agents. Experimental results demonstrate that EvolveRouter significantly outperforms state-of-the-art routing methods across five question-answering benchmarks, achieving notable improvements in both F1 score and exact match metrics, thereby validating the efficacy of co-evolutionary optimization and adaptive collaboration.

Abstract

Large language model agents often exhibit complementary strengths, making routing a promising approach for multi-agent question answering. However, existing routing methods remain limited in two important ways: they typically optimize over a fixed pool of agents without improving the agents themselves, and they often rely on rigid collaboration schemes that cannot adapt the number of participating agents to the query. We propose EvolveRouter, a trainable framework that addresses both limitations by jointly improving agent quality and collaboration structure. First, EvolveRouter couples graph-based query routing with targeted instruction refinement in a closed-loop co-evolution process, allowing router diagnostics to guide agent improvement while refined agents provide cleaner supervision for routing. Second, it introduces an adaptive inference strategy that dynamically determines the effective collaboration size for each query through router-weighted answer agreement. Together, these designs enable more capable and more efficient multi-agent reasoning. Experiments on five question answering benchmarks show that EvolveRouter consistently outperforms SOTA routing baselines in both F1 and exact match, while further analysis confirms the benefits of closed-loop refinement and adaptive collaboration.

Problem

• multi-agent question answering
• routing
• agent collaboration
• adaptive inference
• instruction refinement

Innovation

• co-evolution
• adaptive collaboration
• graph-based routing
• instruction refinement
• multi-agent question answering