摘要
重点介绍了单纯形法在求解过程中常遇到的几种特殊情况.首先,在一个线性规划问题的最优解对应的单纯形表中,如果至少有一个非基变量的检验数为零,那么该线性规划问题的最优解可能不只一个,当求到另一个最优解时,则原问题必有多重最优解;其次,在单纯形表中,如果某一负检验数所对应的列向量的分量全部非正,则原问题无最优解;再次,在求解过程中,若原问题不可行,而对偶问题可行时,我们可以应用对偶单纯形法进行求解.
This paper introduce the basic idea of the simplex algorithm for the standard form of LP,we begin with a basic feasible solution.We always meet the following several cases.Case 1:If there are at least one nonbasis variables;whose relative costs are equal to zero,then the optimal solution of the LP may not be only one while finding another one,the LP maybe have infinite optimal solution.Case 2:If one of the negative relative costs corresponding to components of the column vectors are all nonpositive,then the LP dose not exist optimal solution.Case 3:If the primal LP is not feasible,while its dual is feasible,then we can use the Dual Simplex Algorithm to resolve them.
出处
《甘肃联合大学学报(自然科学版)》
2010年第3期101-103,共3页
Journal of Gansu Lianhe University :Natural Sciences
关键词
单纯形法
对偶单纯形法
换基迭代
最优解
检验数
simplex algorithm
dual simplex algorithm
optimal solution
relative costs