报告题目：Existence of compact sets in neural network control of nonlinear systems

报告人：  向峥嵘 教授  南京理工大学

时  间： 2021年5月7日 10:00-11:00  

地  点：  金沙集团wwW3354CC 专家接待室（305）

报告简介：Neural network is an effective tool for approximating unknown nonlinear functions, and it is widely used in controller design of nonlinear systems. In order to ensure the effectiveness of function approximation, the state variables in the unknown function need to be limited to a compact set. However, in most recent studies, the existence of the compact set is a problem to be avoided, which makes the relevant conclusions have some theoretical defects. Although neural network control scheme has been used to solve many practical engineering problems, its reliability and development are always limited by this theoretical defect. In this talk, this problem will be discussed, and take two typical nonlinear systems as examples to design neural network controller and analyze the stability of closed-loop system.

个人简介：向峥嵘，博士、教授、博士生导师。1998年12月毕业于南京理工大学控制理论与控制工程专业获得博士学位，1998年起在南京理工大学自动化学院工作至今。主持及参与多项国家级与省部级科研项目，获得国防科技进步二等奖，在国内外重要期刊及会议上发表多篇高水平学术论文。担任多项国家级和省级科技项目、科技奖评审专家和多个国内外期刊及会议的审稿人。目前主要从事非线性系统，切换控制系统，多智能体系统等方面的理论及应用研究。