基于模糊加权最小二乘的多粒度语言决策信息集成

Aggregating Multi-granularity Decision Making Information Based on Fuzzy Weighted Least Squares Method

运用最小二乘原理,考虑不同决策专家对其决策结果的信心度并作为信息集成的计算权重,以最终集成结果与单一专家决策信息间的距离平方和为优化目标,建立了多粒度决策信息集成的模糊加权最小二乘规划模型,并以多粒度决策信息的变化范围为约束,与线性距离规划模型进行对比优化计算.结果表明,该模型适用于多粒度语言信息的集成,并能够使最终的群体决策结果尽可能逼近每个决策专家的意见,即降低决策过程中的主观性、计算复杂性及结果残差.通过在水平定向钻机产品方案决策中的实例应用,验证了其具有较好实用性和有效性.

A fuzzy weighted least squares programming model was proposed for aggregating multi-granularity linguistic judgments based on the principle of least squares. In the model, the confidence levels of experts are provided as the weights of the model calculation. The weighted sum of distance squares between the final result and each judgment is used as the optimizing objective. The change ranges of jugements are employed as the constraints of the model. Contrasted to the linear distance programming model, the model can be applicable for the aggregating of multi-granularity linguistic informations and make the final group judgment as close to each individual judgment as possible. The comparson of the results demonstrates the model can decrease the subjectivity, calculating complexity and remaining error in the decision process. The effectiveness and practicability of the proposed model were illustrated and demonstrated in a real-world application of product design altenrative selection for horizontal directional drilling machine.