当前位置: 首页 > 科学研究 > 学术活动 > 正文
刘新为、刘亚锋、蔡邢菊、杨俊锋的学术报告(7.6下午)
发布时间:2022-07-01 00:00  作者:   初审:  复审:  来源:本站原创  浏览次数:

变分分析与优化前沿论坛(7月6日下午)

学术报告(一)

报告题目:A new augmented Lagrangian method for inequality constrained optimization

报告人刘新为 (河北工业大学)

报告时间:2022年7月6日14:30-

腾讯会议:394-2491-2514 密码123123

参加人员:本科生、研究生、教师

报告摘要:We introduce a twice differentiable augmented Lagrangian and present a novel augmented Lagrangian method of multipliers for optimization with general inequality constraints. Our method is a combination of the augmented Lagrangian and the logarithmic-barrier technique, and is a generalization of the Hestenes-Powell augmented Lagrangian. Without assuming any constraint qualification, it is proved that our method has strong global convergence. Locally, our method is capable of rapidly detecting the possible infeasibility, and has linearly convergence to the KKT point. The preliminary numerical experiments on some small benchmark test problems demonstrate our theoretical results.

报告人简介:刘新为教授,博导,1998年博士毕业于中国科学院计算数学与科学工程计算研究所,师从袁亚湘院士。现任河北工业大学数学研究院常务副院长,兼任河北工业大学校学术委员会委员、理学院学术委员会主任,中国运筹学会常务理事、中国运筹学会数学规划分会常务理事,中国数学会计算数学分会理事,河北省运筹学会副理事长兼秘书长,河北省数学会计算数学分会理事长,英文SCI刊物《Mathematical Methods of Operations Research》、《Pacific Journal of Optimization》和《计算数学》编委。先后主持5项国家自然科学基金面上项目、参与1项国家自然科学基金重大研究计划项目。主要研究非凸非线性优化算法及其收敛性理论,在《Mathematical Programming》、《SIAM Journal on Optimization》、《SIAM Journal on Scientific Computing》、《Mathematics of Computation》及《IEEE Transactions on Neural Networks and Learning Systems》等国际重要刊物发表多篇论文。

学术报告(

报告题目:Uplink-Downlink Duality in Wireless Communications: Where Lagrange Meets Shannon(从优化的视角看无线通信中的上下行对偶:一场拉格朗日和香农的对话)

报告人刘亚锋 (中国科学院数学与系统科学研究院)

报告时间:2022年7月6日15:10-

腾讯会议:394-2491-2514 密码123123

参加人员:本科生、研究生、教师

报告摘要:Many problems arising from communication system design can be formulated as optimization problems. In practice, one is often interested in not only the numerical solution to the problems but also the special structure of their optimal solution. In this talk, we shall use some examples from wireless communications and information theory to show that exploring the Lagrangian dual of these (convex) problems often reveal the structure of their optimal solution and the structure of the optimal solution will further lead to better algorithms for solving the corresponding problems.

报告人简介:刘亚锋,2007年毕业于西安电子科技大学理学院数学系,2012年在中国科学院数学与系统科学研究院获得博士学位(导师:戴彧虹研究员);博士期间,受中国科学院数学与系统科学研究院资助访问明尼苏达大学一年(合作导师:罗智泉院士)。博士毕业后,他一直在中国科学院数学与系统科学研究院计算数学所工作,2018年晋升为数学与系统科学研究院副研究员。他的主要研究兴趣是最优化理论与算法及其在信号处理和无线通信等领域中的应用。曾获2011年国际通信大会“最佳论文奖”,2018年数学与系统科学研究院“陈景润未来之星”,2018年中国运筹学会“青年科技奖”,2020年IEEE通信学会亚太地区“杰出青年学者奖”等。他目前担任《IEEE Transactions on Wireless Communications》、《IEEE Signal Processing Letters》和《Journal of Global Optimization》期刊的编委。他是IEEE信号处理学会SPCOM(Signal Processing for Communications and Networking)的技术委员会成员。他的工作获得国家自然科学基金委青年基金、面上项目和优秀青年基金的资助。

学术报告(

报告题目:An balanced Douglas-Rachford splitting algorithm for convex minimization

报告人蔡邢菊(南京师范大学)

报告时间:2022年7月6日16:00-

腾讯会议:394-2491-2514 密码123123

参加人员:本科生、研究生、教师

报告摘要:The Douglas-Rachford algorithm is a classical and effective splitting method to solve the inclusion problems. Recently, an adaptive Douglas-Rachford splitting algorithm is proposed for the monotone inclusion, which allow one operator be weakly monotone. We apply the idea of adaptive Douglas-Rachford splitting method (ADRSM) to differentiable convex optimization problems with abstract constraints, and more attractive results can be obtained for the convex optimization problem. We propose accurate and inaccurate versions of the algorithm respectively, and prove the global convergence of the algorithms. We extend these results to two separable convex optimization problems with linear constraints. In numerical experiments, we compare our algorithms with other commonly used algorithms and the results verify the effectiveness of our algorithms.

报告人简介:蔡邢菊,南京师范大学教授。主要从事最优化理论与算法、变分不等式、数值优化方向。发表SCI论文三十余篇。主持国家青年基金、面上基金、省青年基金各一项,参加国家重点项目一项,获江苏省科技进步奖一等奖一项。担任中国运筹学会副秘书长、江苏省运筹学会秘书长。

学术报告(

报告题目:Primal-Dual Splitting Methods Constructed Based on Convex Combination

报告人杨俊锋 (南京大学)

报告时间:2022年7月6日16:40-

腾讯会议:394-2491-2514 密码123123

参加人员:本科生、研究生、教师

报告摘要:Recently, we have proposed using convex combination technique to construct new primal-dual full splitting algorithms for solving some structured convex optimization problems. In this talk, I will review the proposed algorithms and their convergence properties. A connection to the primal-dual algorithm of Chambolle and Pock will also be given. Recently, we have proposed a golden ratio primal dual algorithm (GRPDA) for solving structured convex optimization problems and it can be viewed as a new adaptation of the classical Arrow-Hurwicz method using a convex combination technique. The convex combination technique has been applied to several related problems.

报告人简介:杨俊锋是南京大学数学系教授、博导。先后师从中国科学院袁亚湘院士、南京大学何炳生教授、美国莱斯大学张寅教授。主要从事最优化计算方法及其应用研究,在SIAM J Optim, Comput Optim Appl等权威期刊发表数十篇论文,开发图像去模糊软件包FTVd、压缩感知L1模解码软件包YALL1等。曾获国家优秀青年基金、中国运筹学会青年科技奖,入选教育部新世纪优秀人才支持计划。