多目标随机决策分析的IRGSD—FDⅡ方法

Multiobjective Stochastic Decision Analysis—IRGSD-FD Ⅱ

本文就多目标随机决策问题,通过引进一种新的解的概念——目标满足度β与约束可行度α协调解,简称(α,β)一协调解。提出了一种决策方法——交互式参考目标满足度与可行度方法Ⅱ(IRGSD—FDⅡ),并讨论了这一方法在研究开发(R&D)项目选择问题中的应用。这一方法的基本步骤包括:替代问题的产生;决策者初始约束可行度α、目标满足度β及目标参考水平q的确定;反映决策者偏好的多属性效用函数的估计;替代目标函数的重新定义和求解;与决策者的对话;最佳(α,β)协调解的获得等等。

By introducing a new solution concept with respect to the multiobjective stochastic decision-making problem-(α,β)composite solution,this paper proposes a new decision analysis method,i.e.,Reference Goal Satisfied Degree-Feasible Degree method Ⅱ(IRGSD-FDⅡ),with a discussion of its applications in R & D Project Selection Problems.The IRGSD-FDⅡ method consists mainly of the following steps:formulation of the surrogate problem:determination of decision maker's initial constraint feasble degree α,goal satisfied degree β and objective's reference level q;evaluation of the mul- ti-attribute utility function reflecting decision maker's preference;redefinition and resolu- tion of the surrogate objective function;interaction with the decision maker:and obtainment of the optimal composite solution.