摘要
We establish in this paper optimal parametric Lagrangian dual models for box constrained quadratic program based on the generalized D.C.(difference between convex) optimization approach,which can be reformulated as semidefinite programming problems.As an application,we propose new valid linear constraints for rank-one relaxation.
We establish in this paper optimal parametric Lagrangian dual models for box constrained quadratic program based on the generalized D.C. （difference between convex） optimization approach, which can be refor- mulated as semidefinite programming problems. As an application, we propose new valid linear constraints for rank-one relaxation.
基金
supported by National Natural Science Foundation of China(Grant Nos. 11001006 and 91130019/A011702)
the Fund of State Key Laboratory of Software Development Environment (Grant No. SKLSDE-2011ZX-15.)
