Some numerical experiments in solving linear programming approximately using sGS-ADMM and its acceleration
报告人：Defeng Sun
报告人单位：The Hong Kong Polytechnic University
时间：22年6月16日（周四）上午 10：00~11：30
线上腾讯会议号：840-962-050 密码：610064
摘要：In this talk, we will present some numerical experiments with the symmetric
Gauss-Seidel based ADMM (sGS-ADMM) and its acceleration for finding approximate solutions to large-scale linear programming (LP) problems. For the purpose of comparison, several other ADMM-type methods and their accelerations are also tested on the same real data sets. From the numerical results, we find that sGS-ADMM and its acceleration are more efficient than other methods and their accelerations. Interestingly, the performance of the sGS-ADMM itself is even faster than some other accelerated methods of the similar nature. Consequently, the sGS-ADMM and its acceleration are suitable in providing an initial solution to warm-start some high-order methods such as the simplex method for handling huge-scale LP problems.
报告人简介：孙德锋，香港理工大学应用数学系系主任和应用优化与运筹学讲座教授，美国工业与应用数学学会会士，中国工业与应用数学学会会士，香港数学学会会长。荣获国际数学规划Beale--Orchard-Hays奖，新加坡国立大学科学学院首届杰出科学家奖。曾任《Asia-Pacific Journal of Operational Research（亚太运筹学杂志）》主编，现任国际顶级数学期刊《Mathematical Programming(数学规划)》编委，《SIAM Journal on Optimization》编委等。在Mathematics of Operations Research, Mathematical Programming, SIAM Journal on Optimization等国际权威刊物上发表学术论文百余篇。主要从事连续优化及机器学习的研究，包括基础理论、算法及应用。在半光滑和光滑化牛顿方法，以及线性和非线性矩阵优化等方面具有很深造诣。其在非对称矩阵优化问题方面取得的系列成果促成了矩阵优化这一新研究方向。 2021年凭借排产方面优化求解器的贡献， 获得华为诺亚方舟实验室杰出合作奖。
邀请人：宋恩彬教授
