Submarine cable network is one of the most important connectivity infrastructures in the digital era.In the past 20 years,the submarine cable network of Chinese mainland has formed a complex connectivity structure.Thi...Submarine cable network is one of the most important connectivity infrastructures in the digital era.In the past 20 years,the submarine cable network of Chinese mainland has formed a complex connectivity structure.This paper focuses on exploring the structure and evolution of the submarine cable network of Chinese mainland.The results show that the evolution can be divided into four stages:an initial stage(1993-1998),a developmental stage(1999-2002),a stagnation stage(2003-2015)and an accelerated stage(2016-2018).The connectivity structure can be analyzed at micro,meso and macro scales.Statistically,the connectivity increased significantly overall,but showed significant differences in space.For the microscale,the landing cities were characterized by“extensive but low,exclusive and high”;for the mesoscale,the connectivity of countries or regions was characterized by“distance attenuation”as a whole,but,in part,by a“regional identity”;for the macroscale,intercontinental connectivity differences have been declining.The hierarchy has been upgraded from a“3 system”to a“2+3 system”.Finally,this paper discusses the interaction between submarine cable network construction and international relations,and puts forward policy suggestions for China’s submarine cable construction.展开更多
为快速构建并准确预测温度作用引起的斜拉桥主梁应变用于结构状态评估,基于某大跨度斜拉桥主梁超过1年的温度和应变监测数据,提出了一种基于迁移学习和双向长短时记忆(bi-directional long short-term memory,Bi-LSTM)神经网络的斜拉桥...为快速构建并准确预测温度作用引起的斜拉桥主梁应变用于结构状态评估,基于某大跨度斜拉桥主梁超过1年的温度和应变监测数据,提出了一种基于迁移学习和双向长短时记忆(bi-directional long short-term memory,Bi-LSTM)神经网络的斜拉桥温度-应变映射模型建立方法。首先,利用解析模态分解(analytical mode decomposition,AMD)去噪应变数据,得到仅由温度引起的应变响应;其次,选择温度和某一测点应变数据构成数据集,采用Bi-LSTM神经网络训练该数据集,并通过网络结构和超参数优化建立温度-应变Bi-LSTM基准模型;最后,利用迁移学习方法,将已训练好的基准模型中部分参数迁移到其他温度-应变数据集,建立相应的温度-应变映射被迁移模型,并与未采用迁移学习的神经网络训练方法进行对比。研究结果表明,相比直接建立的温度-应变Bi-LSTM神经网络映射模型,采用迁移学习方法建立的被迁移模型,其拟合精度均高于所用的基准模型,且训练时间短,预测误差小。展开更多
基金National Natural Science Foundation of China,No.42071151Strategic Priority Research Program of the Chinese Academy of Sciences,No.XDA20010101。
文摘Submarine cable network is one of the most important connectivity infrastructures in the digital era.In the past 20 years,the submarine cable network of Chinese mainland has formed a complex connectivity structure.This paper focuses on exploring the structure and evolution of the submarine cable network of Chinese mainland.The results show that the evolution can be divided into four stages:an initial stage(1993-1998),a developmental stage(1999-2002),a stagnation stage(2003-2015)and an accelerated stage(2016-2018).The connectivity structure can be analyzed at micro,meso and macro scales.Statistically,the connectivity increased significantly overall,but showed significant differences in space.For the microscale,the landing cities were characterized by“extensive but low,exclusive and high”;for the mesoscale,the connectivity of countries or regions was characterized by“distance attenuation”as a whole,but,in part,by a“regional identity”;for the macroscale,intercontinental connectivity differences have been declining.The hierarchy has been upgraded from a“3 system”to a“2+3 system”.Finally,this paper discusses the interaction between submarine cable network construction and international relations,and puts forward policy suggestions for China’s submarine cable construction.
