Experimental validation of domain-level gradient-based and hierarchical PCE-based routing and control in heterogeneous optical networks

A domain-level gradient-based routing （DLR） algorithm for heterogeneous optical networks with syn- chronous digital hierarchy and optical transport network domains is proposed and experimentally vali- dated. This algorithm classifies domains into groups with incremental levels on the basis of domain-level partitioning, and guides paths level by level along a gradient on the basis of interdomain routing tree evolution. The proposed algorithm is implemented in the hierarchical path computation element-based control architecture for connection provisioning. Testbeds with commercial and emulated nodes are es- tablished to verify the feasibility and performance of the algorithm. Experimental and emulation results show that DLR effectively performs in terms of network blocking probability, real time characteristics, and scalability.     

Autoregression Model of Time Series with Matrix Cross-Section Data北大核心CSCD

矩阵型截面数据时间序列的优点在于可以同时刻画多个对象的多个属性.本文重点研究了矩阵型截面数据时间序列的自回归模型,给出了该模型的参数估计、模型识别、白噪声检验三个方面的理论结果.最后再利用矩阵型截面数据时间序列自回归模型,对两支银行股的日收益率序列和日成交量变化率序列进行建模分析.