Journal of Complex Networks

Journal of Complex Networks is the network-science title of Oxford University Press, founded in 2013 and issued six times a year from Oxford. It carries original research and reviews on networks whose topology and dynamics are nontrivial, treating such networks as the skeletons of complex systems, and its 2025 Journal Impact Factor is 1.9, Q2 in mathematics, interdisciplinary applications at 64 of 137.《复杂网络杂志》是牛津大学出版社旗下的网络科学刊物，2013 年创刊，每年在牛津出六期。版面收原创研究与综述，对象是拓扑与动力学都不平凡的各类网络，把它们看作现实复杂系统的骨架；2025 年影响因子 1.9，数学跨学科应用类目列 Q2，排名 64/137。

Two features distinguish it from the general run of applied-mathematics journals. The objects of study are empirical as often as formal, since the networks come from biology, engineering and social life, and the annual output is small, 54 papers in the 2025 count against 1,388 citations to the whole run, which leaves a cited half-life of 6.1 years.该刊与应用数学类期刊的一般面貌有两处不同。研究对象常常来自实证，生物学、工程与社会生活中的网络占了相当篇幅；年发文量不大，2025 年统计为 54 篇，全刊累计被引 1388 次，引用半衰期 6.1 年。

Scope

• The stated aim is a forum for work that advances the analysis and understanding of complex networks and their use in other fields, with networks defined loosely as systems of nontrivial topology and dynamics.宗旨写明要办成论坛，推动对复杂网络的分析与理解，也推动它在其他领域的使用；网络在这里取宽泛定义，指拓扑与动力学都不平凡的系统。
• Mathematics comes first in the remit, from graph theory and spectral methods through random graph models to the numerical analysis that makes large computations possible.数学在范围里排在首位，从图论与谱方法，经随机图模型，一直到支撑大规模计算的数值分析。
• Network theory supplies the working vocabulary, with centrality, clustering, degree distributions, assortativity, community structure and the many motif counts that describe how a graph is put together.网络理论提供基本词汇，中心性、聚类、度分布、同配性、社团结构，以及描述图如何拼起来的各种模体计数，都在其中。
• Dynamics on networks form a second strand, covering synchronisation, diffusion, random walks, coupled oscillators and the spread of perturbations through a connected structure.网络上的动力学自成一脉，同步、扩散、随机游走、耦合振子，以及扰动在连通结构里的传播都归入这一脉。
• Spreading processes and percolation are treated together, so epidemic models, information cascades, robustness under node or edge removal and the fragility of real infrastructures sit in one body of work.传播过程与渗流合在一处处理，流行病模型、信息级联、节点或连边移除下的韧性，以及真实基础设施的脆弱性，构成同一批文献。
• Multilayer and temporal networks have their own place, with coupled layers, changing adjacency, temporal motifs and the methods needed when a single static graph will not do.多层与时间网络另有位置，耦合层、随时间变化的邻接、时序模体，以及单张静态图不够用时的种种方法，都在这里讨论。
• Statistical inference is named in the scope, which brings in network reconstruction from partial data, null models, goodness-of-fit tests and the estimation of generative mechanisms.统计推断写进了范围，从部分数据重建网络、零模型、拟合优度检验，以及生成机制的估计，因此都有落点。
• Social networks appear as a standing subject, from friendship and collaboration structures to the diffusion of behaviour and the measurement problems that come with self-reported ties.社会网络是常设题目，从友谊与合著结构，到行为的扩散，再到自报关系带来的测量难题。
• Biological networks cover metabolic and gene regulatory graphs, protein interaction maps, neural connectomes and food webs, usually with an eye to what the topology implies for function.生物网络覆盖代谢与基因调控图、蛋白质相互作用图、神经连接组与食物网，落笔时多半关心拓扑对功能意味着什么。
• Technological networks take in the internet, transport systems, power grids and communication infrastructure, where the questions are reliability, capacity and failure.技术网络包括互联网、交通系统、电网与通信基础设施，问题多落在可靠性、容量与故障上。
• Informational and ecological systems close the list, with citation graphs, online platforms, species interaction webs and the flows of matter or attention that such structures carry.信息与生态系统收束这份清单，引文图、在线平台、物种相互作用网，以及这些结构承载的物质流与注意力流都在其中。
• The stated endpoint is a predictive model rather than a description, so a paper on molecular, ecological or engineering systems is expected to say what can be forecast and with what error.宗旨给出的落点是可预测的模型，而非一份描述，分子、生态或工程系统上的论文因此要说清能预测什么、误差多大。
• Methods and software belong here too, from community-detection algorithms and embedding techniques to the code and data sets that let a result be reproduced by others.方法与软件同样在范围之内，社团检测算法、嵌入技术，以及让结果能被他人复现的代码与数据集，都算合格内容。
• Pure graph theory with no bearing on real networks sits outside, and so does work that applies a standard tool to a new data set without adding anything to the method.与实际网络无关的纯图论落在范围之外；把一个现成工具套到新数据集上而不改进方法的稿件，同样不受欢迎。

History

• Volume 1 is dated 2013, and the title was set up by Oxford University Press at its Oxford office on Great Clarendon Street, where the editorial address still sits.第 1 卷标为 2013 年，刊物由牛津大学出版社在其克拉伦登街的牛津本部创办，编辑部地址至今仍在那里。
• The launch owed something to timing. Network science had grown for more than a decade after the small-world and scale-free papers of the late 1990s, yet the field had no journal of its own at the meeting point of mathematics and application.创刊与时机有关。1990 年代末的小世界与无标度论文之后，网络科学积累了十余年，数学与应用的交界处却还没有一份自己的刊物。
• Ernesto Estrada, a mathematical chemist who moved into network science, took the editorship and has held it since, running a journal that reads as mathematics while drawing its problems from elsewhere.（Ernesto Estrada）由数学化学转入网络科学，出任主编并在此后一直担任，把刊物办成一副数学刊物的样子，题目却从别处取来。
• The issue count grew with the field. Volume 1 carried two numbers in 2013, volumes 2 and 3 four each and volume 4 three, with six numbers a volume the settled pattern from volume 5 in 2016, which puts the 2025 volume at 13 and the 2026 volume at 14.期数随领域一同成长，第 1 卷在 2013 年出两期，第 2 与第 3 卷各四期，第 4 卷三期，自 2016 年第 5 卷起定为每年六期，2025 年即第 13 卷，2026 年为第 14 卷。
• Volume 8 spans 2019 and 2020, volume 9 sits inside 2021, and the issues of volume 10 carry dates on both sides of that new year, the mark of a title that posts papers online before they enter an issue.第 8 卷跨 2019 与 2020 两年，第 9 卷落在 2021 年之内，第 10 卷各期的标注时间分处那个新年两侧，说明论文先在线发布、再归入期号。
• The web address has moved. Papers from the early volumes point to comnet.oxfordjournals.org, and the current site is academic.oup.com/comnet, the platform on which the whole Oxford list now sits.网址换过。早期卷次的论文指向 comnet.oxfordjournals.org，现行站点是 academic.oup.com/comnet，牛津全套期刊如今都放在这个平台上。
• Access is hybrid. Subscriptions carry the journal, authors may pay for open access, and the gold share reached 24.32 per cent in the 2025 count, a high figure for a title that has never converted wholesale.获取方式是混合型，订阅构成基本盘，作者可选择付费开放，2025 年统计的金色开放占比为 24.32%，一份从未整体转为开放的刊物能有这个比例，并不常见。
• Indexing has been stable. Science Citation Index Expanded covers the journal from volume 1, and PubMed takes selected citations only, so biological and clinical readers meet it through a partial feed.索引情况稳定，科学引文索引扩展版从第 1 卷起收录，PubMed 只取部分引文，生物与临床读者见到的是一份不完整的题录。
• The impact factor has moved in a narrow band, 2.011 in 2021, 1.492 in 2022, 2.2 in 2023, 1.5 in 2024 and 1.9 in 2025, a wobble with no trend strong enough to call either a rise or a decline.影响因子在窄幅内摆动，2021 年 2.011，2022 年 1.492，2023 年 2.2，2024 年 1.5，2025 年 1.9，涨落都不足以称作上升或下滑。
• The five-year figure of 2.4 sits above the 2025 impact factor of 1.9, which means the older papers earn more per article than the recent ones do.五年影响因子 2.4 高于 2025 年的 1.9，说明旧文单篇获得的引用多于新文。
• Output is deliberately small. The 2025 count records 54 articles, 1,388 citations to the whole run and a cited half-life of 6.1 years, figures that fit six issues of roughly nine papers each.发文量有意压得小，2025 年统计为 54 篇，全刊累计被引 1388 次，引用半衰期 6.1 年，与每年六期、每期约九篇的容量相合。
• Bibliographic reporting lags the title. The national library still files the frequency as two issues a year, a label that fitted only the first volume and has not been corrected.书目记录落后于刊物本身，国家医学图书馆的频率栏仍写作一年两期，这个标注只在第 1 卷成立，也没有改过。
• One JCR category is assigned, mathematics, interdisciplinary applications, where the 2025 rank is 64 of 137 and the quartile Q2, while the Journal Citation Indicator stands at 0.53 and the percentile at 53.6.期刊引证报告只给了一个类目，即数学跨学科应用，2025 年排名 64/137，分区 Q2，期刊引文指标 0.53，百分位 53.6。

Editors

• Ernesto Estrada of the Institute of Interdisciplinary Physics and Complex Systems in Palma edits the journal, a post he has held since the founding volume.主编是帕尔马交叉物理与复杂系统研究所的（Ernesto Estrada），自创刊卷起担任此职。
• His own work runs from spectral graph theory to the walkability and communicability of networks, and his 2011 monograph The Structure of Complex Networks set out much of the vocabulary the journal now uses.他本人的研究从谱图论延伸到网络的可游走性与可通信性，2011 年的专著《The Structure of Complex Networks》为刊物如今使用的许多术语立了框架。
• The board is drawn from mathematics, physics, computer science and biology, a spread that matches a remit running from proof to data set.编委从数学、物理学、计算机科学与生物学中选取，这一分布与从证明到数据集的刊稿范围相配。
• That spread does real work, because the citation reports file the journal under a single mathematics heading while a good share of what it prints is judged by physicists and biologists.这一分布有实际用处，引证报告只把它归入一个数学类目，而它刊出的稿件有相当部分要由物理学家与生物学家来评判。
• Two article types carry the journal, research papers and reviews, and a review is expected to map a whole direction rather than summarise the output of one group.刊物主要有两类稿件，研究论文与综述，综述要为一个方向画出全图，而非总结一个课题组的产出。
• Authors are asked for economy. The guidance puts a research paper at roughly seven to eight thousand words including figures and references, and asks that the argument be clear on a first reading.指南要求文字经济，研究论文连图表与参考文献在内以七八千词为宜，论证要让人一读就明白。
• Readership is mixed, so notation has to be introduced rather than assumed, and a paper on contagion in a contact network must stay legible to the epidemiologists who will use it.读者成分混杂，符号须逐一交代而不能默认读者熟悉，一篇讨论接触网络传染的论文，仍要让会用它的流行病学家读得下去。
• Referees come from the board and from outside it, and a paper that crosses from method to application usually needs a report from each side before a decision is reached.评审人取自编委，也取自编委之外，一篇从方法跨到应用的稿件，通常要两侧各有报告才能定夺。
• The editorial test is stated in the aims. An application paper has to end in a model that predicts something, and a method paper has to change what can be computed or inferred.编辑标准写在宗旨里，应用类论文要落到能预测的模型上，方法类论文要改变可计算或可推断的边界。
• Production, hosting and the submission system belong to the publisher, which leaves the editor free to judge the science and keeps the journal's address at Oxford.制作、托管与投稿系统由出版社承担，主编因此只需判断学术内容，刊物的地址也一直留在牛津。