报告题目:On Solution of Constrained Tikhonov Regularization via Modulus-Based Iteration Methods

报告人：郑宁 特任研究员

报告时间：上午9:00—10:00

报告地点：广知 A 楼 214

博学堂讲座预告

报告题目：On Solution of Constrained Tikhonov Regularization via Modulus-Based Iteration Methods

报告时间：2019.05.04  (周六) 上午9:00—10:00

报告地点：广知 A 楼 214

报告人：郑宁 特任研究员

摘要： Tikhonov regularization is commonly applied to the solution of linear discrete ill-posed problems. In many applications the desired solution is known to be nonnegative. It is then natural to require that the approximate solution determined by Tikhonov regularization is also nonnegative. The present paper describes iterative methods, based on modulus methods, for computing nonnegative approximate solutions of Tikhonov regularization problems in general form. Methods suitable for small and large problems are discussed. Computed examples illustrate the performances of the methods considered.

本报告为综述性介绍报告，欢迎广大本科生、研究生前来参加。

报告人简介：

郑宁，男，同济大学应用数学专业学士，同济大学应用数学博士，日本国立信息学研究所基础信息学特任研究员。现任日本理化学研究所智能创新研究中心特别研究员。研究方向是数值代数、科学计算、张量和机器学习。在SIAM J. Matrix Anal. Appl., J. Comput. Appl. Math., J. Sci. Comput., Numer. Algorithms等国际专业杂志发表论文十余篇，且有数篇论文在投。多次参加国际会议并给出会议报告。