摘要
在文献［１］以枢轴运算为基础的算法基础上，引入基向量的成本和非基向量的偏差等概念并将后者也纳入枢轴运算范畴，另外介绍具有上下界线性函数的处理方法。
This paper presents a pivoting-based algorithm called projective algorithm in which the basic units processed are sperate constraints or their coefficient vectors, row vectors, in contrast to the column vectors in simplex algorithm where all the general constraints are bound together in the form of standard linear programming (SLP).The two algorithms are equivalent in the sense that when the SLP is solved by simplex algorithm,the dual of SLP is sovled by the other. However,the new algorithm need not add any variable to the original problem or replace a free variable by two nonnegative ones. Since the independence of all constraints is maintained,redundant constraints may be found and eliminated in the iterations. This algorithm needs less computation and less memory than simplex algorithm to solve any problem.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
1996年第3期316-320,共5页
Journal of University of Electronic Science and Technology of China
关键词
枢轴运算
单纯形算法
投影算法
线性规划
pivoting operation
simplex algorithm
Projective algorithm
basic cone
