方案优选属于多属性决策问题。不确定语言多属性决策是一种研究与应用较多的多属性决策方法,决策的重点是对不确定语言变量的集结算法。决策过程中,需要对不确定语言变量进行对比排序。针对已有的排序方法存在误差的情况,给出了一种能够准确计算可能度的模型,该模型满足不确定语言变量的下标取值服从均匀分布的条件。在不确定语言变量的集结算法中,ULHGM算子适用性强,但它先考虑了数据自身的重要性程度,然后才考虑数据位置的重要性程度,这在一定程度上会造成评价信息的失真,影响集结结果的准确性。为此,对其集结过程进行了改进,根据赋予算子关联权重的初衷,直接对专家的原始评判数据赋予位置权重,然后赋予反映数据自身重要性程度的权重,最后再进行集结。ULHGM算子需要进行两次指数加权运算才能得出集结结果,而改进后的算子只需要进行一次指数加权运算,因此其集结算法更为简便。接下来,基于ULWGM算子和改进的ULHGM算子,建立了方案优选的模型。最后,以某舰炮维修性设计方案的评选为例,分别用IULHGM算子和ULHGM算子对专家的评价信息进行集结,得出了不同的方案排序结果,并对原因进行了分析。案例表明IULHGM集结算法更加准确、合理。

Scheme selection is a Multiple Attribute Decision Making (MADM) problem. The Uncertain Linguistic MADM (ULMADM) method is generally studied and widely used. The key point for this kind of decision making is the aggregation algorithm. It is necessary to compare and rank uncertain linguistic variables in the process of decision making. To overcome the error occuring in the existed ranking model, a new model that can accurately calculate the possibility degree is presented, which guarantees that the subscript value of an uncertain linguistic variable is subject to uniform distribution. During the aggregation algorithms, the uncertain linguistic hybrid geometric mean (ULHGM) operator is of good applicability. This operator deals with the weight of a datum itself ahead of the position weight, so it might allocate counterproductive position weights to experts' original opinions. To avoid this, its aggregation process is improved, and it directly allocates position weights to experts' original opinions. The ULHGM operator needs twice exponential weight operations to complete aggregation, but the improved ULHGM (IULHGM) operator only needs once, so the improved algorithm is more convenient. Then the scheme selection method based on ULWGM operator and IULHGM operator is formulated. Finally, taking the selection of the maintainability design scheme of a certain shipborne gun as an example, the experts' evaluation data are aggregated by IULHGM operator and ULHGM operator, respectively, which lead to different scheme ranking results, and the cause is analyzed. The case shows that the aggregation algorithm of IULHGM operator is more accurate and reasonable than that of ULHGM operator.