A Feasible Direction Algorithm for Max Bisection via Low-RankFactorization
图的最大二等分问题的低秩可行方向算法

Based on the semidefinite programming relaxation of {\rm max} Bisection, a feasible direction algorithm is given to solve the relaxation. Coupled with the 0.699 randomized method, the approximate solution of {\rm max} Bisection is obtained. Furthermore, its convergent result is given. The numerical experiment shows that the algorithm can solve the {\rm max} Bisection efficiently.