摘要
对大规模的线性规划问题,确定最优解需要大量的计算机存储空间和运算时间.如果在求解大规模线性规划问题之前,先对其进行数据预处理,化约约束条件,减少约束矩阵非零元素,提高稀疏度,则可达到缩小问题规模,节省计算机内存和运算时间的效果.本文系统地讨论线性规划问题的数据预处理方法,分别对常变量、零变量、非极点变量以及约束矩阵多余非零元素进行理论分析。
In the field of production administration and management decision as well as in many other fields, people depend more and more on the method of Linear Programming to make the optimum decisions. There are many algorithms used to solve Linear Programming problems, but it is possible to get half the result with twice the effort if some direct algorithms are used. Especially it needs a great deal of computers memory and calculating time to determine the optimum solution to the largescale problem of Linear Programming, which costs a lot. Therefore, if we make the data processing in advance, simplify its constraint conditions and decrease the nonzero element in the constraint matrix before solving the largescale Linear Programming problem, then we can achieve the result of reducing the scale of the problem, saving the computer memory and computing time. We systematically discuss the method of data processing in advance, analyzing theoretically the constant variable, null variable, nonextremity variabl
出处
《沈阳化工学院学报》
1999年第2期83-86,共4页
Journal of Shenyang Institute of Chemical Technolgy
关键词
线性规划
数据预处理
规划论
linear programming
\ constant variable
\ null variable
\ nonextremity variable
\=nonzero elemente and the redundancy nonzero element of the constraint matrix respectively and presenting the processing method.\=