Mei_et_al-2016-srep26649.pdf                      刚发表在ScientificReports的文章介绍

众所周知，网络结构决定其功能。目前大量研究的是网络结构如何决定和影响网络功能, 反过来的问题这些年也有大量研究, 但还远远不够。我们在基于自适应同步的网络结构识别方面已经取得一定进展[1-11], 但更实际的问题是基于实际数据的网络结构反演, 最近几年开始已有一些研究工作，比如基于格兰杰因果检验的方法[12-13]，由噪声诱导的用动力学相关性判断网络结构[14]，及基于优化和压缩感知的结构识别等[15-16]。这些方法都取得了很好的效果，并且在进一步地发展和应用。上述结果均是针对结构不变的网络。 

然而网络并非静止而是时刻变化的。这一变化包括网络结构的变化，网络所承载的流的变化以及网络要素属性和功能的变化。事实上，对大多数我们感兴趣的复杂系统，比如疾病传播网络和社会关系网络，网络结构的动态演化特性都至关重要。这些动态网络中，节点间的相互作用随着时间的变迁而动态变化，或仅在某一小段或几段时间内保持不变，并不像在静态网络中那样始终处于稳定不变的状态。这种动态特征传统的静态网络无法直接表示。 

ScientificReports的这篇文章《Identifying structures ofcontinuously-varying weighted networks》针对的是结构连续时变的稀疏复杂网络， 从纯数据出发，提出基于优化的方法成功实现网络结构的动态识别[17]。同时， 由于引入了正则化， 我们的方法有较好的抗噪性。另外，由于该算法只需要少量观测数据，从而降低了实际应用中数据测量的难度。在50000个节点的 small-world network 成功地识别连续时变的边的权重。该方法的提出，使得我们在网络结构识别方面又向更广泛的实际应用迈进了一步。

[1].      Jin Zhou ,and Jun-an Lu，Topologyidentification of weighted complex dynamical network, PhysicaA386(2007): 481-491.

[2].      XiaoqunWu，Synchronization-basedtopologyidentification of weighted general complex dynamical networks withtime-varyingcoupling delay , Physica A  387(2008):997-1008.

[3].       HuiLiu, Jun-an Lu, JinhuLü, andDavid John Hill，Structure identification of uncertain generalcomplex dynamicalnetworks with time delay，Automatica45(2009):1799-1807

[4].       Liang Chen, Jun-an Lu, and Chi K. Tse,Synchronization: An Obstacle toIdentification of Network Topology,IEEE Transactions onCircuitsandSystems-II 56(2009):310-314.

[5].    Jin Zhou, Wenwu Yu, Xiumin Li, Michael Small, and Jun-anLu，Identifying the Topology of aCoupledFitzHugh–NagumoNeurobiological Network via a Pinning Mechanism，IEEE Transactions on Neural Networks 20(2009):1679-1684.

[6].      JunchanZhao, Qin Li,Jun-An Lu andZhong-Ping Jiang,Topology Identification of complex dynamicalnetworks, Chaos 20(2010):023119.

[7].    Longkun Tang , Jun-an Lu, Xiaoqun Wu , JinhuLü, Impact of nodedynamics parameters on topology identification of complex dynamical networks, NonlinearDynamics 73(2013): 1081-1097.

[8].    Shuna Zhang, Xiaoqun Wu,  Jun-an Lu, Hui Feng,JinhuLü,Recovering structures of complex dynamical networks based ongeneralized outersynchronization, IEEE Transactions on Circuits and Systems-I 61 (2014):3216-3224.

[9].   Xiaoqun Wu, Xueyi Zhao, JinhuLü, Longkun Tang, Jun-an Lu,Identifying topologies of complex dynamical networks with stochasticperturbations,   IEEE Transactionson Control of Network Systems,   doi:10.1109/TCNS.2015.2482178, 2015.

[10].Yingfei Wang,Xiaoqun Wu, Hui Feng, Jun-an Lu, JinhuLü, Inferring topologies and detectinghidden sources of complex dynamical networks with perturbations, Science China E, doi:10.1007/s11431-016-6050-1, 2016.

[11].Yingfei Wang,Xiaoqun Wu, Hui Feng, Jun-an Lu, Yuhua Xu, Inferring topologies viadriving-based generalized synchronization of two-layer networks,   Journal of Statistical Mechanics 2016 (2016):053208.

[12]. XiaoqunWu, Changsong Zhou, GuanrongChen,and Jun-an Lu, Detecting thetopologies ofcomplex networks with stochastic perturbations, Chaos 21(2011), 043129.

[13]. Xiaoqun Wu, Weihan Wang, andWei Xing Zheng, Inferringtopologies of complex networks with hidden variables,Phys. Rev. E 86 (2012): 046106.

[14].JuanChen, Jun-an Lu, Jin Zhou, Topologyidentification of complexnetworks from noisy time series using ROC curveanalysis, Nonlinear Dynamics 75(2014): 761–768.

[15].Tao He,Xiliang Lu, Xiaoqun Wu, Jun-an Lu,Wei Xing Zheng,Optimization-based structure identification of dynamical networks, PhysicaA 392(2013):1038-1049.

[16].Guanjun Li, Xiaoqun Wu, Juan Liu, Jun-an Lu, ChiGuo, Recoveringnetwork topologies via Taylor expansion and compressive sensing,  Chaos 25 (2015): 043102.

[17].Guofeng Mei, Xiaoqun Wu, Guanrong Chen, Jun-an Lu, Identifying  structures of continuously-varying weightednetworks,  ScientificReports 6 (2016): 26649.www.nature.com/articles/srep26649