In this paper, unconstrained minimization with block-circulant structured polynomials is studied. A specifically designed method is presented to show that it can solve problems with sizes much larger than the classica...
详细信息
In this paper, unconstrained minimization with block-circulant structured polynomials is studied. A specifically designed method is presented to show that it can solve problems with sizes much larger than the classical Lasserre's semidefinite relaxation. The proposed approach is in the same spirit of Lasserre's relaxation but with a careful exploration of the underlying circulant structure, which helps reducing the sizes of the result semidefinite program problems significantly. Despite of the reduction, a certification for the global optimality is derived as well.
暂无评论