摘要
讨论离散半无限规划问题,结合更新离散指标集的技术,提出一种新的可行序列二次规划(FSQP)算法求解由半无限规划(SIP)转化到离散半无限(DSI)问题,使得可行下降方向仅通过求解一个QP子问题可获得,为克服马太效应,高阶校正通过求解带有包含某个约束集的线性方程组所得。在适当的条件下,证明了算法的全局收敛性和超线性收敛性。
A class of finely discretized semi-infinite programming （SIP） problems are discussed in this paper. In light of the technique of updating discretization index set, we present a new feasible sequential quadratic programming （FSQP）＇algorithm to solve the Discretized Semi-Infinite （DSI） problems from SIP. A feasible descention is obtained by solving only one QP sub-problem. In order to avoid Maratos effect, a high-order revised direction is computed by solving a linear system with involving some ＂active＂ constraints. The theoretical analysis shows that weak global and superlinear convergence can be deduced.
出处
《桂林电子科技大学学报》
2009年第2期122-125,共4页
Journal of Guilin University of Electronic Technology
基金
国家自然科学基金(10501009)
广西自然科学基金(桂科自0728206)
中国博士后科学基金(20070410228)
关键词
半无限规划
可行序列二次规划
线性方程组
全局收敛性
超线性收敛性
semi-infinite programming （SIP） problems
feasible sequential quadratic programming （FSQP）
linear system
global convergence
superlinear convergence
