题目一:Semidefinite Relaxations for MIMO Detection: Tightness, Tighterness, and Beyond
内容简介:Multiple-input multi-output (MIMO) detection is a fundamental problem in modern digital communications. Semidefinite relaxation (SDR) based algorithms are a popular class of approaches to solving the problem because the algorithms have a polynomial-time worst-case complexity and generally can achieve a good detection error rate performance. In this talk, we shall first develop two new SDRs for MIMO detection and show their tightness under an easily checkable condition. This result answers an open question posed by So in 2010. Then, we shall briefly talk about the tighterness relationship between some existing SDRs for the MIMO detection problem in the literature. Finally, if time is allowed, we shall also talk about a branch-and-bound algorithm (based on the newly derived SDR) for globally solving the MIMO detection problem (and a more general class of nonconvex complex quadratic problems).
报告人:中国科学院数学与系统科学研究院计算数学所 刘亚锋 副研究员
报告人简介: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罗智泉教授一年。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,已在Mathematical Programming, SIAM Journal on Optimization, Mathematics of Operations Research等优化期刊以及 IEEE Transactions on Signal Processing, IEEE Journal on Selected Areas in Communications, IEEE Transactions on Wireless Communications, IEEE Transactions on Information Theory等IEEE交叉领域期刊发表论文三十余篇。曾获2011年国际通信大会“最佳论文奖”(由IEEE通信学会颁发),2015年WiOpt (International Symposium on Modeling and Optimization in Mobile, Ad Hoc and Wireless Networks)“最佳学生论文奖”,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”等。他目前担任《IEEE Transactions on Wireless Communications》和《IEEE Signal Processing Letters》期刊的编委和《Journal of Global Optimization》期刊的客座编委。他是IEEE高级会员(Senior Member)、亚太信号与信息处理学会(Asia-Pacific Signal and Information Processing Association)无线通信和网络(Wireless Communications and Networking)方向的技术委员会成员(Technical Committee)、中国运筹学会数学规划分会副秘书长。
题目二:Semidefinite Relaxations for MIMO Detection: Tightness, Tighterness, and Beyond Some new parallel algorithm for the eigenvalue and SVD problems
内容简介:HSS matrix is an important kind of rank-structured matrix, which can be used for solving integral equations, eigenvalue and SVD problems. In this talk we will introduce how to use it to accelerate the DC algorithm for the bidiagonal SVD problem and the tridiagonal eigenvalue problems. The proposed algorithm will be denoted by ADC. When dealing with large matrices with few deflations, ADC can be 3x faster than DC in the optimized LAPACK libraries such as Intel MKL without any degradation in accuracy. We further show another SVD algorithm based on Zolotarev’s Polar Decomposition algorithm, which is highly scalable. When implemented on parallel computers, it can be two times faster than the DC algorithm in ScaLAPACK. We will show some results obtained on Tianhe 2 Supercomputer.
报告人:国防科技大学计算机学院计算机所 李胜国 助理研究员
报告人简介:计算数学博士,现为国防科技大学计算机学院计算机所助理研究员,2006年国防科技大学理学院应用数学专业本科毕业,2008年和2013年获得国防科技大学的计算数学硕士和博士学位。2010-2012年再美国加州大学伯克利分校联合培养两年,师从Gu Ming教授。2013年底参加工作以来,主要从事并行算法设计、特征值计算、共性算法库研制、Benchmark程序测试与优化工作,参与天河2A、银河-X研制与系统调试工作,主持国家青年和湖南省面上自然科学基金各1项,曾荣获湖南省优秀硕士论文和全军优秀博士论文,发表SCI论文20多篇,两篇进入ESI检索前10%,部分论文发表在SIAM J. Sci. Comput., SIAM J. Matrix Anal. Appl., Numer. Math., Numer. Linear Algebra Appl. Parallel Computing等。
时 间:2019年9月17日(周二)上午9:30始
地 点:南海楼124室
热烈欢迎广大师生参加!
网络空间安全学院
2019年9月16日