A STRONGLY CONVERGENT METHOD OF QUASI-STRONGLY SUB-FEASIBLE DIRECTIONS FOR CONSTRAINED OPTIMIZATION

作者：Jian, Jin-Bao^*; Chen, Yue-Hua; Guo, Chuan-Hao

来源：Pacific Journal of Optimization, 2011, 7(2): 339-351.

• constrained optimization
• method of quasi-strongly sub-feasible directions
• global convergence
• strong convergence

In this paper, based on the method of strongly sub-feasible directions and linear programming approach, a new method for constrained optimization problems, called the method of quasi-strongly sub-feasible directions, is presented. The main features of this new method are as follows: (1) the starting point can be chosen arbitrarily; (2) the operations of initialization (Phase I) and optimization (Phase II) can be unified automatically without using any penalty parameters; (3) the number of functions satisfying the constrained conditions is nondecreasing in the iterative procedure; (4) by introducing a new non-monotone line search and a slightly new technique for computing the search direction, the proposed algorithm possesses global and strong convergence (i.e., the whole sequence of iterative points converges to a KKT point). Some preliminary numerical results are also reported.

• 出版日期2011-5
• 单位南宁师范大学; 广西大学; 上海大学