多目标线性规划的无目标淹没交互式决策方法(NOSIA)

A NEW INTERACTIVE METHOD OF MULTIOBJECTIVE LINEAR PROGRAMMING

本文从实用决策角度出发,分析了现有MOLP所存在的目标淹没问题和应用的某些局限性。以凸集理论为基础,提出目标一致性量度矩阵、基准解集和无目标淹没搜索概念,给出求解决策支持矩阵的有效算法。并进一步建立了无目标淹没搜索(NOSS)模型,给出求解MOLP问题的无目标淹没的交互式决策方法(NOSIDM),该方法能够准确方便地反映决策者偏好作用,便于交互作用决策。

In this paper objective submerged problems and limitations existing in the methods currently used for multiobjective linear programming (MOLP) are analyzed related to real-world processess. The concepts of consistency measure matrix, basical reference solution set and non-objective-submerging are presented on the basis of cenvex set theory, and a fast algorithm of solving the matrix is given. Furthermore, non-objective-submerged searching model is built according to the principle of equal satisfying integration of objectives. Finally, the non-objective-submerged and interactive approach is implemented, where non-objective—submerging exists and exactly refleets the aspiration or the prcferences of decision makers.