数学科学学院学术报告
Application and Innovation: Deploying Novel Adaptive Dynamic Zeroing Neural Networks for Solving Time-Varying Generalized Sylvester Equations
张彦钧
（重庆理工大学）
报告时间：2024年9月14日 星期六 上午9：00-10：00
报告地点：腾讯会议：429-536-693 会议密码：0914
报告摘要：The time-varying generalized Sylvester equation, essential in areas like control theory, signal processing, and system identification, can be effectively converted into a series of linear equations. Traditional zeroing neural network (ZNN) models, while delivering effective results, often grapple with the challenge of balancing computational complexity and system robustness. To address these issues, this paper introduces two improved ZNN models that significantly enhance computational efficiency and adaptability to dynamic environmental changes. These improvements are realized through the integration of a new strong dynamic adaptive coefficient algorithm and the introduction of two innovative activation functions that feature variable exponents. Comprehensive theoretical analyses coupled with extensive numerical experiments clearly demonstrate these models' enhanced capabilities in effectively solving the time-varying generalized Sylvester equation. The findings of this study significantly enhance the application potential of ZNN in complex industrial environments. In addition, they offer new insights for future optimization of this model.
报告人简介：张彦钧，硕士生导师，2023年入职重庆理工大学理学院。研究方向包括随机算法，统计计算和神经网络。在NLAA, LAA, JCAM等期刊发表论文多篇，主持国家自然科学基金和重庆市教委青年项目。
邀请人：谢家新
