Home > Scientific Research > Paper Publications

AN ALTERNATING DIRECTION METHOD FOR SOLVING A CLASS OF INVERSE SEMI-DEFINITE QUADRATIC PROGRAMMING PROBLEMS

Hits:

Indexed by:Journal Papers
Journal:JOURNAL OF INDUSTRIAL AND MANAGEMENT OPTIMIZATION
Included Journals:SCIE、Scopus
Volume:12
Issue:1
Page Number:317-336
ISSN No.:1547-5816
Key Words:Inverse semi-definite programming; alternating direction method; spectral projected gradient method; BB-step; convex semi-definite quadratic programming
Abstract:In this paper, we propose an alternating-direction-type numerical method to solve a class of inverse semi-definite quadratic programming problems. An explicit solution to one direction subproblem is given and the other direction subproblem is proved to be a convex quadratic programming problem over positive semi-definite symmetric matrix cone. We design a spectral projected gradient method for solving the quadratic matrix optimization problem and demonstrate its convergence. Numerical experiments illustrate that our method can solve inverse semi-definite quadratic programming problems efficiently.
Date of Publication:2016-01-01

Pre One:Probability approximation schemes for stochastic programs with distributionally robust second-order dominance constraints

Next One:Convergence analysis on a smoothing approach to joint chance constrained programs