双目标规划问题的像集与求解
2016-08-11马赞甫刘妍珺
马赞甫 刘妍珺

摘要经济管理的决策目标往往与成本、收益相关,双目标规划在经济管理中具有广泛应用.然而,尚缺乏成熟的算法确定双目标规划问题的全部解.给出双目标规划问题像集的一般性确定法,以求其解,为研究目的所在.具体而言,构造一个带等式约束的单目标规划问题,以确定双目标规划问题像集之部分边界,并借助拉格朗日乘子符号判断其单调性,据此确定原问题的帕累托解与弱帕累托解.这相当于提供了一个求解双目标规划问题的一般性框架.
关键词最优化;双目标规划解法;单目标规划;像集;弱帕累托解
中图分类号F224-3 文献标识码A
AbstractThe objective of decisionmaking in economic management is often related to cost and benefit. A good case in point is that biobjective programming based on costbenefit analysis is widely used in economic management. However, so far, there is still a lack of mature algorithms to determine the full solution of the Bi objective programming problem. A general method, which was used to get the Pareto solution or weak Pareto solution for double objective optimization problems, was put forward in this research. To be specific, we constructed a single objective programming with equality constrain to determine the frontier of the image set of double objective optimization problems. Furthermore, we could determine the frontier's functional monotonicity using Lagrange multiplier and finally get the Pareto solution or weak Pareto solution. Base on above steps, A general framework was presented to solve double objective programming problem.
Key wordsOptimization, The Method of Solving DoubleObjective Programming, SingleObjective Programming, Image Set, Weak Pareto Solution
1引言
多目标规划的解法可区分为间接算法与直接算法两种.间接法的一般思路是将多个目标转换为单一目标进行处理,往往只能得到问题的一个解或部分解.如杨轶华、吕显瑞、刘庆怀等(2006)所指出的,间接法存在顾此失彼之缺点[1].直接法则以同伦法为代表,通过构造同伦映射确定多目标规划问题的解,直指多目标规划问题本身,可回避间接法之缺点.其中,刘庆怀、林正华(2000)考察采用该算法确定问题的最小弱有效解[2],而杨轶华、吕显瑞、刘庆怀等(2006)则采用该算法确定多目标规划问题的一个有效解集[1].
考虑到万事万物好与坏、收益与成本等双分类的普遍性,多目标规划中的双目标规划具有极为重要的应用价值,常用于企业管理决策[3-5]、交通设置优化[6-7]等现实问题.不失一般性……
