线性规划初始对偶可行基本解的一种求法  

A Method for Achieving an Initial Dual Feasible Basic Solution for Linear Programming

作　　者：李蕊[1] 王艳红[1] 

机构地区：[1]西安工业大学理学院,西安710021

出　　处：《西安工业大学学报》2014年第3期173-176,共4页Journal of Xi’an Technological University

摘　　要：运用对偶单纯形法求解线性规划问题时,需要先给定一个初始对偶可行的基本解.然而在线性规划问题的约束条件Ax=b中,矩阵A一般不含m阶单位矩阵,此时初始对偶可行的基本解不易求得.文中通过对线性规划问题增加人工变量和一个约束条件,给出一步便能求出其初始对偶可行基本解的简便方法,进而通过对偶单纯形法进行迭代解决线性规划问题.His necessary to give an initial dual feasible basic solution ,when solving linear programming problems by the dual simplex method .However ,the matrix A of the linear programming constraint Ax= b does not usually contain the identity matrix .So ,the initial dual feasible basic solution is difficult to obtain directly .In this paper ,by adding artificial variables and a constraint ,an initial dual feasible basic solution is obtained at one step ,and then the optimal solution of the original problem is obtained by iterations .

关 键 词：两阶段法 对偶单纯形法 扩充问题 对偶可行基本解 

分 类 号：O221.1[理学—运筹学与控制论]