非凸二次规划的分支定界方法

A Branch and Bound Algorithm for Nonconvex Quadratic Programming

通过构造二次函数的线性下界函数给出非凸二次约束二次规划问题(QP)的松弛线性规划,提出分支定界算法,数值计算表明算法是有效可行的。

First, a linear relaxation program ot the nonconvex quadratically-eonstramed quadrane programming （QP） is given by constructing a lower linear function of quadratic functions. A branch and bound algorithm is then proposed for the problem （QP）. Finally a numerical computation is given to illustrate the feasibility and efficiency of the proposed algorithm.