复杂系统模糊模式识别动态规划模型研究

Fuzzy recognition dynamic programming model for complex system decisions

将模糊识别集合算子和非结构决策分析方法引入模糊动态规划中,推导出了两种求解复杂系统多目标多阶段决策问题的方法:级别特征值最小法和阶段模糊识别递推法。这两种方法具有不同的递推方程,可以应用于不同的情况,从而形成了模糊模式识别动态规划模型。此模型可以有效地获取和集成决策者的经验知识和偏好。此模型已应用到实践中,并取得良好效果。

This paper develops a fuzzy recognition dynamic programming model for hybrid multiobjective multistage decision making problems arising from complex systems A fuzzy aggregation operator is first derived from a fuzzy pattern recognition model to get a global evaluation for all objectives Then, by incorporating the operator into the traditional dynamic programming, two dynamic approaches are proposed to solve the problems, ie, the minimum ranking approach and the stage recognition approach, which form the fuzzy recognition dynamic programming model The two approaches are of characteristics of different back recurrence equations A nonstructural decision analysis method is simultaneously used to compute the objective weights in the dynamic process of decision makings Finally, a case study is given to show the model is reasonable and feasible in practice