多目标广义分式规划的Lagrange最忧性条件和对偶理论  

MULTIOBJECTIVE GENERALIZED FRACTIONAL PROGRAMMING:LAGRANGE OPTIMALITY CONDITIONS AND DUALITY THEORY

作　　者：李锋[1] 

机构地区：[1]云南师范大学数学系

出　　处：《云南师范大学学报（自然科学版）》1996年第2期1-10,共10页Journal of Yunnan Normal University:Natural Sciences Edition

摘　　要：本文考虑多目标广义分式规划（ＶＰ），其中目标函数定义为任意多个分式函数的逐点最大值。对规划（ＶＰ），引入向量值Ｌａｇｒａｎｇｅ函数，讨论了最优性条件和鞍点条件；引入带参数的多目标规划（ＶＰｅ），证明了鞍点存在性定理；证明了规划（ＶＰ）的Ｌａｇｒａｎｇｅ对偶理论和Ｗｏｌｆｅ对偶理论。In this paper, we consider multiobkctive generalized fractional programming(VP), in which objective functions are pointwise maximum of finite or infinite ratio functions. For programming(VP), we definite its vector-valued Lagrangian, discuss its optimality conditions and saddle-point theorem, prove its saddle-point cxistence theorem,lagrange duality theory and wolfe duality theory, via multiobjcctivc parametric programming(VPe).

关 键 词：分式规划 对偶理论 多目标规划 最优性条件 

分 类 号：O221[理学—运筹学与控制论]