二分单纯形算法中子规划的改进

二分单纯形算法中，线性规划问题的最优解是通过求解一系列子问题来实现的，本文针对二分单纯形算法中的子规划问题作进一步研究，提出了一个新的了规划问题来改善问题的不可行性，并确定出了相应的主元旋转规则，给出了相应的子算法，同时进行了数值实验，实验结果表明，调用新子算的二分法与原始二分法相比，迭代次数和计算时间均有所改善，可视为原始二分算法的一种改进算法。

An Improvement on the Sub-Program of the Simplex-Like Method with Bisection

In order to solve linear programming problems, simplex method with bisection carries out pivoting operation on a series of sub-program.Further study on the sub-program is made in this paper,and a new special linear sub-program is introduced to improve the infeasibility of the problem. The authors describe the corresponding pivoting rule,set up the algorithms, and carry out numerical test in a set of linear programming problems.The results of the test indicate that compared with the original one,the algorithm using the new sub-program reduces the numbers of iterations and running time by quite large margins. So,it can be regarded as an improved algorithm to the original one.