Adaptive estimation of parameters and delays in dynamical models

发布者:系统管理员发布时间:2011-04-19浏览次数:1984

报告题目: Adaptive estimation of parameters and delays in dynamical models
报 告 人: 林伟
  复旦大学数学科学学院教授,副院长
报告时间: 4月22日上午10:00
报告地点: 数学系第一报告厅
相关介绍: Abstract: In real applications, with observed data, dynamical models, whose concrete form can be either known or unknown, need establishing for data verification and prediction.  For the model forms that are, a priori, established by the knowledge of physical, biological, and chemical knowledge, parameters and delays in the models can be estimated by the adaptive synchronization technique under some mild but necessary conditions.  However, ordinary parameters estimation is somewhat different from the time delays estimation in dynamical models.   As for the case of unknown models, the adaptive synchronization technique combined with the radial basis function neural network can be imported to establish model forms in a practical sense through the observed data.   Concrete examples of biological and physical significance are provided to show the practical usefulness of the proposed techniques.  In particular, unstable periodic orbit is approximately located only through the data without an, a priori, established model. 
XML 地图 | Sitemap 地图