Autonomous Agents and Multi-Agent Systems

Autonomous Agents and Multi-Agent Systems is the official journal of the International Foundation for Autonomous Agents and Multi-Agent Systems, founded in 1998 and published by Springer, covering the theory, engineering and application of agents and multi-agent systems, with a 2025 Journal Impact Factor of 2.4.《自主智能体与多智能体系统》是国际自主智能体与多智能体系统基金会（International Foundation for Autonomous Agents and Multi-Agent Systems）的官方会刊，1998 年创刊，由 Springer 出版，内容覆盖智能体与多智能体系统的理论、工程与应用，2025 年期刊影响因子为 2.4。

Scope

• The stated remit is to disseminate significant original results on the foundations, theory, development, analysis and applications of autonomous agents and multi-agent systems, and the editors keep that remit broad.公开定位是发表自主智能体与多智能体系统的重要原创成果，基础理论与开发方法都在其中，应用分析同样在内，编辑部把这一范围放得很宽。

• Agent decision-making architectures form the first block, covering deliberative and practical reasoning, reactive or behavioural designs, plan-based control and hybrid systems, together with the evaluation of each architecture.智能体决策架构是第一个板块，覆盖慎思式推理与实践推理，反应式或行为式设计也在其中，基于计划的控制与混合架构同样受理，各类架构的评估方法一并纳入。

• Cooperation and teamwork follow, with organisational structuring and design for multi-agent systems, self-organisation, emergent functionality and swarm intelligence all treated as core subjects in their own right.合作与团队协作紧随其后，多智能体系统的组织结构与设计列为核心课题，自组织与涌现功能在列，群体智能同样被当作独立课题对待。

• Multi-agent planning belongs here as well, including the coordination of plans and activities across agents, and so do distributed constraint processing and distributed constraint optimisation.多智能体规划同样属于这一范围，涉及多个智能体之间计划与活动的协调，分布式约束处理与分布式约束优化也在此列。

• A large computational economics strand takes in auction and market systems, algorithmic mechanism design, automated negotiation, cooperative and non-cooperative game theory, and computational social choice.计算经济学是分量很重的一条线，拍卖与市场系统、算法机制设计都在其中，自动谈判与博弈论同样受理，合作与非合作两个分支一并覆盖，计算社会选择理论也归入这一条。

• Knowledge representation and reasoning, together with the logical foundations of autonomous agents, are treated as a subject in their own right rather than as a preliminary to building systems.知识表示与推理，以及自主智能体的逻辑基础，被当作独立课题对待，没有被降为建造系统的预备步骤。

• Agent programming languages are assessed on implementation, semantics and evaluation, so a paper may argue about language design as readily as it reports a working system.智能体编程语言从实现层面与语义层面加以衡量，评估结果同样算数，论文既可以讨论语言设计，也可以报告一套可运行的系统。

• Multi-agent argumentation and dialogue are in scope, and so are the detection and resolution of conflict between agents that disagree about facts, plans or norms.多智能体论辩与对话在收稿范围之内，智能体之间围绕事实、计划或规范产生的冲突，其检测与消解同样受理。

• Learning runs through the whole remit, from computational architectures for learning agents and their evolution and adaptation to multi-agent learning, co-learning and evolutionary approaches.学习贯穿整份收稿范围，从学习型智能体的计算架构及其演化与适应，一直延伸到多智能体学习与协同学习，演化方法也在其中。

• Agent communication languages appear with their semantics, pragmatics and implementation, alongside communication protocols and the conversations those protocols make possible.智能体通信语言连同语义与语用层面一并受理，实现问题同样在内，通信协议以及协议支撑的对话过程也在其中。

• Conventions, commitments, norms, obligations and social laws are studied as coordination devices, and models of trust and reputation sit in the same group.惯例与承诺被当作协调手段来研究，规范与义务同样如此，社会法则也在这一组，信任模型与声誉模型归入同一处。

• Believable and synthetic agents and characters belong to the remit too, with human-agent interaction treated as a research problem in its own right.可信的合成智能体与合成角色同样在范围内，人机交互被当作独立的研究问题处理。

• Environments, testbeds and programming languages built for experiments form a section of their own, and robotic agents bring integrated perception, cognition and action, cognitive robotics, robot planning and multi-robot systems.为实验而建的环境与测试床自成一类，配套的编程语言也在其中，机器人智能体另成一类，涵盖感知与认知及行动的集成，也涵盖认知机器人学与机器人规划，多机器人系统同样受理。

• Ontologies for agent systems, agents on the semantic web, grid-based systems and service-oriented computing complete the architectural side of the list.面向智能体系统的本体与语义网上的智能体各自成题，网格计算系统与服务导向计算补全了架构一侧的清单。

• Agents also count as a software engineering paradigm, which brings specification and verification, methodologies, comparative evaluations and reports of practical experience into the journal.智能体还被当作一种软件工程范式，规格说明与验证由此进入刊物，方法论与不同路线的比较评估同样受理，实践经验报告也在其列。

• The scope reaches psychology, linguistics, philosophy and human-computer interaction on one side, economics, sociology and organisation theory on the other, and it closes with novel applications and authoritative reviews and tutorials.范围一侧伸向心理学与语言学，也伸向哲学与人机交互，另一侧伸向经济学与社会学，组织理论同样在内，末尾落在新颖应用以及权威综述与教程上。

History

• The first volume appeared in March 1998, when volume 1 issue 1 opened a run that has continued without interruption to the present day.第 1 卷第 1 期于 1998 年 3 月出版，自此连续出刊至今，中间没有中断。

• Kluwer Academic Publishers brought out the early volumes, a fact still visible in the 10.1023 prefix carried by the 1998 papers, while articles added after the move to Springer carry the 10.1007 prefix.早期各卷由 Kluwer Academic Publishers 出版，1998 年论文 DOI 上的 10.1023 前缀仍能看出这段来历，转入 Springer 之后发表的论文改用 10.1007 前缀。

• The opening issue led with 'A Roadmap of Agent Research and Development', a survey that laid out the research agenda the field would follow for the next decade.创刊号以《A Roadmap of Agent Research and Development》开篇，这篇综述勾画了此后十余年该领域的研究议程。^[1]

• The AAMAS conference that shares the field's name began only in 2002, when AGENTS, ICMAS and ATAL merged, so the journal predates that merger by four years.与本刊同名的 AAMAS 会议到 2002 年才出现，由 AGENTS 和 ICMAS 以及 ATAL 三家会议合并而成，期刊的创刊比这次合并早四年。

• Volume numbering has run ahead of the calendar because some years carried two volumes, so 2019 closed at volume 33 and 2026 opened volume 40.卷号跑在年份前面，有些年份一年出两卷，2019 年收在第 33 卷，2026 年则开到第 40 卷。

• Six issues a year was the rule up to 2019, when volume 33 still ran issues 1 to 6, and since 2020 each volume has carried two issues, one in each half of the year.2019 年及以前每年出 6 期，第 33 卷仍排到第 6 期，2020 年起每卷只出两期，上下半年各一期。

• Springer has published the journal since the Kluwer list was absorbed, and the copyright line now reads Springer Science+Business Media, part of Springer Nature.Kluwer 的刊目并入之后，刊物一直由 Springer 出版，版权行如今写作 Springer Science+Business Media（隶属 Springer Nature）。

• IFAAMAS is a non-profit foundation that promotes science and technology in artificial intelligence, autonomous agents and multi-agent systems, and the journal is its official publication.IFAAMAS 是一家非营利基金会，宗旨是推动人工智能领域以及自主智能体与多智能体系统方向的科学与技术，本刊即其官方出版物。

• The publishing model is hybrid, the journal page counted 210 open access articles at the time of reading, and the 2025 Journal Citation Reports credit 54.29 percent of output to gold open access.出版模式为混合开放获取，读刊页时统计到的开放获取论文有 210 篇，2025 年期刊引证报告把 54.29% 的产出记为金色开放获取。

• The 2025 metrics are a Journal Impact Factor of 2.4 and a five-year value of 2.9, with a second-quartile place at 44 of 88 in automation and control systems.2025 年的指标是影响因子 2.4、五年影响因子 2.9，在自动化与控制系统类目 88 种刊物中列第 44 位，属 Q2。

• In artificial intelligence the journal sits at 139 of 210 in the third quartile, a first decision comes in a median of nine days, and 2025 brought 277.6 thousand downloads.在人工智能类目 210 种刊物中列第 139 位，属 Q3，首次决定的中位时间为 9 天，2025 年全年下载量为 27.76 万次。

• Coverage takes in Science Citation Index Expanded, Scopus, Ei Compendex, INSPEC, DBLP, the ACM Digital Library and Current Contents/Engineering, Computing and Technology.收录范围包括科学引文索引扩展版与 Scopus，也包括 Ei Compendex 和 INSPEC，DBLP 与 ACM Digital Library 同样在列，Current Contents/Engineering, Computing and Technology 一并覆盖。

• Altogether 962 DOIs have been registered, from March 1998 to 27 October 2026 in volume 40 issue 2.登记在册的 DOI 累计 962 条，最早一条来自 1998 年 3 月，最新一条落在 2026 年 10 月 27 日的第 40 卷第 2 期。

Editors

• Yevgeniy Vorobeychik of Washington University in St. Louis and Michael Winikoff of Victoria University of Wellington share the editor-in-chief role.主编由（Yevgeniy Vorobeychik）与（Michael Winikoff）共同担任，前者在圣路易斯华盛顿大学（Washington University in St. Louis），后者在惠灵顿维多利亚大学（Victoria University of Wellington）。

• Vorobeychik works on multi-agent systems, game theory and machine learning, with security games and adversarial machine learning among his better-known lines of research.（Yevgeniy Vorobeychik）研究多智能体系统与博弈论，也研究机器学习，安全博弈与对抗性机器学习是他较受关注的方向。

• Winikoff's subject is agent programming, with declarative agent languages and the verification of agent behaviour at the centre of his work.（Michael Winikoff）的课题是智能体编程，声明式智能体语言与智能体行为的验证处于他工作的中心。

• Fifty-six associate editors, listed alphabetically and each attached to a university or a research institute, support the two editors-in-chief.两位主编之下有 56 位副编辑，名单按姓氏字母顺序排列，每人挂靠一所大学或研究机构。

• The list opens with Stefano Albrecht at Nanyang Technological University and Natasha Alechina at Utrecht, and closes with William Yeoh at Washington University, Makoto Yokoo at Kyushu and Dengji Zhao at ShanghaiTech.名单从南洋理工大学的（Stefano Albrecht）与乌得勒支大学的（Natasha Alechina）起头，排到华盛顿大学的（William Yeoh），再往下是九州大学的（Makoto Yokoo）与上海科技大学的（Dengji Zhao）。

• Mechanism design and computational social choice are represented by Tuomas Sandholm at Carnegie Mellon and Haris Aziz at UNSW Sydney, argumentation by Katie Atkinson at Liverpool and Elizabeth Black at King's College London.机制设计与计算社会选择有卡内基梅隆大学的（Tuomas Sandholm）与新南威尔士大学的（Haris Aziz），论辩方向由利物浦大学的（Katie Atkinson）与伦敦国王学院的（Elizabeth Black）代表。

• Robotics and multi-agent learning bring in Katia Sycara at Carnegie Mellon, Sven Koenig at the University of Southern California and Peter Stone at Texas at Austin, while Munindar Singh at North Carolina State covers normative multi-agent systems.机器人与多智能体学习方向有卡内基梅隆大学的（Katia Sycara），南加州大学的（Sven Koenig）与得克萨斯大学奥斯汀分校的（Peter Stone），北卡罗来纳州立大学的（Munindar Singh）负责规范多智能体系统。

• The associate editors are distributed over eighteen countries and regions, with the United States, Britain and the Netherlands supplying the largest groups and the rest scattered from Israel and Brazil to Japan, China, Hong Kong and New Zealand.副编辑分布在 18 个国家与地区，美国与英国的人数最多，荷兰次之，其余学者散在以色列与巴西，也散在日本与中国香港，新西兰同样有人任职。

• Eight manuscript types are open to authors, namely regular papers, editorials, survey papers, tutorials, reviews, viewpoints, brief communications and book reviews.可投的稿件分八种，常规论文与社论之外还有综述论文与教程，评论与观点文章各占一种，简报与书评补足最后两种。

• Editorials are kept very brief and introduce the papers of a special issue or outline the theme that holds such an issue together.社论篇幅很短，用来引出某一期专刊的论文，或者点明把这一期串起来的主题。

• Survey papers and tutorials are expected to be comprehensive and authoritative, and book reviews are held to the same standard for books about agents and multi-agent systems.综述论文与教程要求内容全面且具有权威性，书评面对智能体与多智能体系统主题的著作时也按同一标准衡量。

• Brief communications exist for important observations that can be stated clearly in a short format, rather than at the length of a full paper.简报为那些能用短篇幅讲清楚的重要观察留出位置，不需要铺成一整篇论文的篇幅。

• Under the arrangement with IFAAMAS, research articles accepted in the twelve months before each year's AAMAS notification date may be presented at the conference, with an optional two-page extended abstract in the proceedings.依据与 IFAAMAS 的约定，每年 AAMAS 录用通知日之前 12 个月内被本刊接收的研究论文可以在会议上报告，会议录里可附一篇不超过两页的扩展摘要，是否刊出由作者选择。