摘要
利用KT条件、罚函数法,将三层线性规划降为约束条件为线性的二层规划,再利用Frank-Wolf线性逼近的理论,从而仅需求解一层线性规划就得到了三层线性规划的最优解。其中线性规划的求解应用了主元标单纯形法,其优点是可以得到更靠近最优点的可行解,从而减少计算量。
KT Condition and the penalty function are used to turn linear three level programming problem into bilevel programming problem whose constraint conditions are linear. Then by using Frank-Wolf linear approximation theory, we obtain the optimal solution of the linear three level programming problem by solving linear programming. The principal pivot simplex method is used to solve linear programming. Its merit is to obtain the feasible solution which is to approach the optimal point, thus reduces the amount of computation.
出处
《湖南工业大学学报》
2009年第1期36-39,共4页
Journal of Hunan University of Technology
关键词
三层线性规划
线性逼近
主元标
单纯形法
罚函数法
three level linear programming
linear approximation
principal pivot
simplex method
penalty function