We propose a new approach for justifying complexity bounds for dual optimization methods. Dual problems often have very big or unbounded size of the optimal solutions. This makes impossible to apply to the complexity analysis of corresponding schemes the standard framework. In this talk we propose new methods, which can work with unbounded feasible sets. All these methods are primal-dual: they generate both primal and dual solutions with required accuracy/feasibility guarantees.
Current 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996