篮彩购买app

Projected Subgradient Methods in InfiniteDimensional Spaces

发布者:文明办发布时间:2019-11-21浏览次数:409


主讲人:Professor Xu, Hong-Kun


时间:2019年12月3日13:00


地点:数学系332室


举办单位:数理学院


主讲人介绍:南非国家科学院院士,发展中国家科学院院士


内容介绍:Subgradient methods, introduced by Shor and developed by Albert, Iusem,Nesterov,  Polyak, Soloov, and many others, are used to solve nondifferentiableoptimization  problems. The major differences from the gradient descent methods(or  projection-gradient methods) for differentiable optimization problems lie in  theselection manners of the step-sizes. For instance, constant step-sizes for  differen-tiable objective functions no longer work for nondifferentiable  objective functions;for the latter case, diminishing step-sizes must however be  adopted.In this talk, we will first review some existing projected subgradient  methodsand the main purpose is to discuss weak and strong convergence of  projected sub-gradient methods in an infinite-dimensional Hilbert space. Some  novel approachesfor strong convergence analysis of projected subgradient methods  will particularlybe presented.