摘要
定义了一种偏离Kuhn-Tucker三元点的度量函数的基础上,对一般连续可微非线性规划提出了一个新的全局收敛算法。利用这个算法在获得问题最优解的同时,还得到了与最优解相应的Lagrange乘子。把这种算法应用于二次规划,得到了二次规划的一种新的迭代法。最后给出了一个计算实例。
A new globally convergent algorithm was presented for a continuous differen-tiable nonlinear programming by defining a measure function deviating from the Kuhn-Tucker point. With this algorithm which can be used to get the optimum solution of the problem,the optimal lagrangian multiplier corresponding the optimal solution of the problem was also obtained. A new iterative algorithm for quadratic programming is obtained when applying the general algorithm to quadratic programming. Finally,a numerical example was given.
出处
《重庆大学学报(自然科学版)》
EI
CAS
CSCD
1993年第5期125-130,共6页
Journal of Chongqing University
关键词
度量函数
全局收敛
非线性规划
measure function deviating from the Kuhn-Tucker triad point
global conver-gence
feasible region
vertical projection