Dynamic Trust-Aware Sparse Communication Topology for LLM-Based Multi-Agent Consensus
2026-06-01 • Multiagent Systems
Multiagent SystemsArtificial Intelligence
AI summaryⓘ
The authors study how groups of language-based agents can work together better by talking only when necessary instead of all the time, which saves time and effort. They propose DySCo, a method that picks the most important conversations based on trust and task relevance, so agents only share useful information. Their approach also ends discussions early when everyone agrees, making the process more efficient. They test DySCo on different reasoning tasks and show it balances communication cost with maintaining good teamwork.
large language modelsmulti-agent systemscommunication topologydynamic sparse consensustrust mechanismcross-validationmulti-round deliberationreasoning taskscommunication complexity
Authors
Wanshuang Gou, Zihan Liu
Abstract
Large language model-driven multi-agent systems enhance the reliability of complex reasoning tasks through multi-round deliberation, role specialization, and cross-validation. However, existing multi-agent debate and collaboration frameworks typically adopt fully connected communication, causing the number of messages, token costs, and end-to-end latency to grow approximately quadratically with the number of agents; although fixed sparse topologies reduce overhead, they cannot adapt communication relationships to different task instances or intermediate reasoning states, making them prone either to preserving low-value interactions or to losing critical error-correction information. To address this problem, this paper proposes DySCo (Dynamic Sparse Consensus), a dynamic trust-aware sparse consensus mechanism. In each round of reasoning, DySCo estimates the value of communication edges based on agent reliability, answer divergence, and task relevance, and selects a small number of high-value edges for message exchange under budget constraints; it then aggregates the answers of different agents through dynamic trust weights and terminates the discussion early once consensus stabilizes. This mechanism replaces universal broadcasting with on-demand communication, thereby reducing communication overhead while preserving essential cross-validation information. We further present analyses of communication complexity and consensus stability, and evaluate the performance of DySCo on mathematical reasoning, logical reasoning, and factual question-answering tasks.