A long-step primal–dual path-following method for semidefinite programming

0.00 Avg rating—0 Votes

Article ID:	iaor20011531
Country:	Netherlands
Volume:	23
Issue:	1/2
Start Page Number:	53
End Page Number:	61
Publication Date:	Aug 1998
Journal:	Operations Research Letters
Authors:	Jiang J.M.
Keywords:	programming: linear

Abstract:

In this paper, the long-step primal–dual path-following method of Jansen et al. is extended from linear programming to semidefinite programming. By generalizing the proximity measure and using the Nesterov–Todd direction as the search direction, we construct an algorithm which finds an ϵ-optimal solution in at most O(n log (ϵ)) iterations.

Reviews

Required fields are marked *. Your email address will not be published.