面向开放环境的分布式多智能体协作框架

doi:10.3969/j.issn.1000-5641.2026.05.004

摘要/Abstract

摘要：

随着大语言模型赋予智能体更强的语义理解与任务执行能力, 多智能体系统正从封闭内网走向开放网络, 在跨机构金融审计、分布式科学计算、去中心化智能服务等场景中展现出广阔前景. 然而, 开放环境中参与者互不信任、利益多元、无中心权威, 导致现有系统的协作过程不可验证、故障难以复现与恢复、行为无法追责. 本文系统梳理了开放环境下多智能体协作面临的核心挑战, 包括状态转换的语义正确性与日志可验证性之间的内在冲突、长期记忆的可信存储与版本一致性问题, 以及自利智能体参与下的贡献归因与激励相容困境. 在此基础上, 提出以区块链与可验证计算为技术基底的理论框架, 包括链上链下协同的总体架构、可验证容错状态机、语义记忆的可验证存储与一致性维护、以及抗操纵的行为归因与激励相容机制. 本研究将推动分布式系统、形式化验证、博弈论与大语言模型不确定性的深度交叉, 为“人工智能+”行动提供可信任协同的基础理论支撑.

关键词: 多智能体, 区块链, 自愈韧性

Abstract:

As large language model (LLM) equips agents with stronger semantic understanding and task execution capabilities, multi-agent systems are transitioning from closed intranets to open networks. They show great potential for cross-institutional financial auditing, distributed scientific computing, decentralized intelligent services, and related scenarios. However, open environments often lack mutual trust, involve heterogeneous interests, and operate without a central authority. As a result, existing systems struggle to support verifiable collaboration, reproducible and recoverable failure handling, and accountable agent behavior. The objective of this study was to systematically review the key challenges of multi-agent collaboration in open environments. These challenges include the tension between the semantic correctness of state transitions and the verifiability of logs, trusted storage and version consistency of long-term memory, and contribution attribution and incentive compatibility among self-interested agents. Based on this analysis, a theoretical framework grounded in blockchain and verifiable computation is proposed. This framework includes an integrated on-/off-chain architecture, verifiable fault-tolerant state machines, verifiable storage and consistency maintenance for semantic memory, manipulation-resistant attribution, and incentive mechanisms. This research will foster deep cross-fertilization among distributed systems, formal verification, game theory, and LLM uncertainty, providing fundamental theoretical support for trustworthy collaboration in the “AI+” initiative.

Key words: multi-agent systems, blockchain, self-healing resilience

中图分类号:

TP181

张召, 李欣然. 面向开放环境的分布式多智能体协作框架[J]. 华东师范大学学报（自然科学版）, 2026, 2026(5): 42-50.

Zhao ZHANG, Xinran LI. Distributed multi-agent collaboration framework for open environments[J]. J* E* C* N* U* N* S*, 2026, 2026(5): 42-50.

图/表 2

参考文献 22

1	Wooldridge M, Jennings N R.. Intelligent agents: theory and practice. The Knowledge Engineering Review, 1995, 10 (2): 115- 152.
2	Sycara K P.. Multiagent systems. Artificial Intelligence Magazine, 1998, 19 (2): 79- 92.
3	Tran K T, Dao D, Nguyen M D, et al. Multi-agent collaboration mechanisms: a survey of LLMs [PP/OL]. V1. arXiv (2025-01-10)[2026-07-11]. https://arxiv.org/abs/2501.06322.
4	Rezazadeh A, Li Z, Lou A, et al. Collaborative memory: multi-user memory sharing in LLM agents with dynamic access control [PP/OL]. V1. arXiv (2025-05-23)[2026-07-11]. https://arxiv.org/abs/2505.18279.
5	Han R, Yan Z, Liang X Q, et al.. How can incentive mechanisms and blockchain benefit with each other? A survey. ACM Computing Surveys, 2023, 55 (7): 1- 38.
6	Hua Y, Chen H, Wang S, et al. Shapley-Coop: credit assignment for emergent cooperation in self-interested LLM agents [C]//Advances in Neural Information Processing Systems 38. 2025: 98517-98544.
7	Wang J H, Zhang Y, Gu Y J, et al. SHAQ: incorporating Shapley value theory into multi-agent Q-learning [C]//Advances in Neural Information Processing Systems 35. 2022: 5941-5954.
8	Fathalla E, Azab M, Xin C S, et al.. Self-sovereign identity as a secure and trustworthy approach to digital identity management: a comprehensive survey. ACM Computing Surveys, 2026, 58 (7): 190.
9	Rebello G A F, Camilo G F, de Souza L A C, et al.. A survey on blockchain scalability: from hardware to layer-two protocols. IEEE Communications Surveys & Tutorials, 2024, 26 (4): 2411- 2458.
10	Heo J W, Ramachandran G S, Dorri A, et al.. Blockchain data storage optimisations: a comprehensive survey. ACM Computing Surveys, 2024, 56 (7): 179.
11	Shi C M, Xie H M, Yan Z, et al.. A survey on off-chain technologies. ACM Computing Surveys, 2026, 58 (7): 181.
12	Wu Q, Bansal G, Zhang J, et al. AutoGen: enabling next-gen LLM applications via multi-agent conversation [PP/OL]. V2. arXiv (2023-10-03)[2026-07-11]. https://arxiv.org/abs/2308.08155.
13	Hong S, Zhuge M, Chen J, et al. MetaGPT: meta programming for a multi-agent collaborative framework [C]//International Conference on Learning Representations. 2024: 23247-23275.
14	Chen B, Li G L, Lin X, et al. BlockAgents: towards Byzantine-robust LLM-based multi-agent coordination via blockchain [C]//ACM Turing Award Celebration Conference 2024. ACM, 2024: 187-192.
15	Chen R N, Dong Y, Liu Y Z, et al. FLock: robust and privacy-preserving federated learning based on practical blockchain state channels [C]//Proceedings of the ACM on Web Conference 2025. ACM, 2025: 884-895.
16	Sun H C, Li J, Zhang H Y. zkLLM: zero knowledge proofs for large language models [C]//Proceedings of the 2024 on ACM SIGSAC Conference on Computer and Communications Security. ACM, 2024: 4405-4419.
17	Lewis P, Perez E, Piktus A, et al.. Retrieval-augmented generation for knowledge-intensive NLP tasks. Advances in Neural Information Processing Systems, 2020, 33, 9459- 9474.
18	Park J S, O’Brien J, Cai C J, et al. Generative agents: interactive simulacra of human behavior [C]//Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology. ACM, 2023: 1-22.
19	Li Z, Xi C, Li C, et al. MemOS: a memory OS for AI system [PP/OL]. V4. arXiv (2025-12-03)[2026-07-11]. https://arxiv.org/abs/2507.03724.
20	Raza S, Sapkota R, Karkee M, et al.. TRiSM for agentic AI: a review of trust, risk, and security management in LLM-based agentic multi-agent systems. AI Open, 2026, 7, 71- 95.
21	Deng Z, Guo Y, Han C, et al.. AI agents under threat: a survey of key security challenges and future pathways. ACM Computing Surveys, 2025, 57 (7): 1- 36.
22	Zhang P Y, Ding S, Zhao Q L.. Exploiting blockchain to make AI trustworthy: a software development lifecycle view. ACM Computing Surveys, 2024, 56 (7): 163.

[1]	金澈清, 周倩, 张召, 杨艳琴. 可扩展区块链架构设计与展望[J]. 华东师范大学学报（自然科学版）, 2026, 2026(5): 26-41.
[2]	纪坤华, 金浩阳, 刘文辉, 黄定江. 大语言模型与多智能体赋能新型电力系统[J]. 华东师范大学学报（自然科学版）, 2026, 2026(5): 178-195.
[3]	乔少杰, 蒋宇河, 刘晨旭, 金澈清, 韩楠, 何帅为. 基于智能合约的教育大数据安全管理和隐私保护算法[J]. 华东师范大学学报（自然科学版）, 2024, 2024(5): 128-140.
[4]	黄超然, 佟兴, 张召, 金澈清, 杨英杰, 秦钢. 面向教育的区块链应用合约架构和数据隐私研究[J]. 华东师范大学学报（自然科学版）, 2022, 2022(5): 61-72.
[5]	杨英杰, 秦钢, 陈淑刚, 尹芙蓉, 张召, 金澈清. 钢材区块链质保书研发及供应链应用[J]. 华东师范大学学报（自然科学版）, 2022, 2022(5): 36-47.
[6]	邓思佳, 佟兴, 唐海波, 张召, 金澈清. 面向区块链的数据管理中间件[J]. 华东师范大学学报（自然科学版）, 2021, 2021(5): 60-73.
[7]	尹芙蓉, 朱承宇, 赵斌, 张召. 基于CITA区块链的纠删码分片存储实现[J]. 华东师范大学学报（自然科学版）, 2021, 2021(5): 48-59.
[8]	肖大龙, 张敬伟, 杨青, 周娅. 基于以太坊的社区贡献激励方法[J]. 华东师范大学学报（自然科学版）, 2020, 2020(6): 82-89.
[9]	黄立波, 王伟, 徐彦军, 陈旭刚. 基于区块链的数字结业证书管理系统及其性能评估[J]. 华东师范大学学报（自然科学版）, 2020, 2020(6): 72-81.
[10]	刘峰, 杨杰, 李志斌, 齐佳音. 一种面向双中台双链架构的内生性数据安全交互协议研究[J]. 华东师范大学学报（自然科学版）, 2020, 2020(5): 44-55.
[11]	高一琛, 赵斌, 张召. 面向以太坊的智能合约自动生成方法研究与实现[J]. 华东师范大学学报（自然科学版）, 2020, 2020(5): 21-32.
[12]	齐学成, 朱燕超, 邵奇峰, 张召, 金澈清. 基于区块链的仓单管理系统[J]. 华东师范大学学报(自然科学版), 2018, 2018(5): 144-153.
[13]	田福粮, 田秀霞, 陈希. 基于区块链的智能电表身份认证方案[J]. 华东师范大学学报(自然科学版), 2018, 2018(5): 135-143,171.