提出了求解不定二次规划问题一个新的分支定界算法。利用D.C.分解和正定阵的Cholesky分解把问题转化为可分离形式,并导出Lagrangi=对偶界,给出基于Lagrangian对偶界和矩形对分的分支定界算法,同时给出初步数值实验结果。
In this paper a new algorithm for finding a global solution of indefinite quadratic programming is proposed. The problem is first refermulated as a separable form by D.C. decomposition and Cholesky factorization. And then the Lagrangian dual bound is derived. A new branch-and-bound algorithm based on the Lagrangian dual bounds and rectangular bisection is presented. Finally, preliminary numerical results are reported.