一类随机多目标二次线性规划模型的交互式算法  

INTERACTIVE ALGORITHM FOR A CLASS OF STOCHASTIC MULTI-OBJECTIVE QUADRATIC-LINEAR PROGRAMMING MODELS

作　　者：徐玖平[1] 李军[1] 

机构地区：[1]四川大学信息与决策研究所,成都610065

出　　处：《系统科学与数学》2002年第1期96-106,共11页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金(70171021)资助课题

摘　　要：针对线性约束条件下带有一个二次目标函数和多个线性目标函数的随机多目标决策问题，借助参考方向法和权重法对该决策问题的期望值模型进行标量化，获得了关于期望值模型的（恰当／弱）有效解的充要条件，引入Ａｃｈｉｅｖｅｍｅｎｔ函数建立了一类随机多目标二次线性规划模型的交互式计算方法．In this paper, a stochastic multi-objective programming model with one qua-dratic and several linear objectives is introduced, which is to be optimized subject to linear constraints. By means of the expectation of random variables, the reference direction approach and the weighted sums approach are used to make the expectation model scalar. At the same time, the sufficient and necessary conditions about (properly, weakly) efficient solutions are obtained and an interactive algorithm is proposed for a class of stochastic multiple objective quadratic linear programming.

关 键 词：多目标决策 交互式算法 有效解 多目标二次线性规划 

分 类 号：O221.1[理学—运筹学与控制论] O221.5[理学—数学]