线性规划的约束条件“滚雪球”算法  

An Algorithm of Restrained Conditions ″Rolling Snowball″ for Linear Programming

作　　者：唐建国[1] 

机构地区：[1]湖南科技学院数学与计算科学系,湖南永州425006

出　　处：《数学的实践与认识》2006年第4期135-143,共9页Mathematics in Practice and Theory

基　　金：湖南省自然科学基金(03JJY3014)资助

摘　　要：为使线性规划的每个约束条件部分或全部地拥有原整个约束条件所包含的信息,将线性规划的约束条件“滚雪球”后得到与原约束条件等价的新约束条件,对新约束条件所构成的线性规划采用目标函数最速递减算法.有一定规模的随机数值算例显示了该算法只需进行m(约束条件数)次迭代即可求得最优解.In order to partly or completely possess the information contained in whole original constraint conditions of linear programming in each constraint condition, new constraint conditions equivalent to original ones are obtained after the constraint conditions of the original linear programming ＂roiling snowball＂. The fast decreasing algorithm of objective function is applied to linear programming with new constraint conditions. Random numerical examples on a small scale show that only m （number of constraint conditions） iterations are needed for finding the optimal solution of linear programming.

关 键 词：线性规划 单纯形方法 对偶单纯形方法 目标函数最速递减算法 约束条件“滚雪球”算法 

分 类 号：O221.1[理学—运筹学与控制论] TP301.6[理学—数学]