摘要
对传统大M法进行改进,若计算检验数的表达式中含有M则只计算含有M的部分,从而简化计算,迭代过程中当人工变量由基变量变为非基变量时,直接去掉人工变量部分的表格然后继续计算,从而再一次降低计算量。借鉴两阶段法的优点进一步给出了无需给出大M的迭代算法,此法不会破坏目标函数的一致性,而且可以避免传统大M法在利用计算机求解时由于M值的选取不当所导致的计算错误。
Improved big-M method is presented. If expressions of the calculated test number contain M, the only portion containing M is calculated, and thereby the calculation is simplified. And when artificial variables become nonbasic variables by basic variables in the iterative calculation process, the artificial variables parts of the table can be directly removed and then the calculation is continued. Thus, the amount of computation is again reduced. Taking advantages of two-phase method, an iteration algorithm without giving the big M is further given. This method does not undermine the consistency of the objective function, and the calculation error can be avoided when using traditional big-M method combined with computer to solve, due to the improper selection of the value of M.
出处
《计算机工程与应用》
CSCD
2014年第18期59-62,69,共5页
Computer Engineering and Applications
基金
安徽省高等学校省级自然科学研究项目(No.KJ2011B152)
关键词
线性规划
单纯形法
大M法
两阶段法
linear programming
simplex method
big-M method
two-phase method
