<![CDATA[Objective weighting methods are essential in multi-criteria decision-making (MCDM) because criterion weights directly influence alternative rankings. However, dispersion-based weighting approaches may assign excessive importance to criteria with high variability, resulting in an imbalanced weighting structure. This study proposes LOGSTA-D, an extension of the Logarithmic Normalization and Standard Deviation (LOGSTA) method that introduces a Dispersion Compression Transformation based on the square root of standard deviation. The transformation is designed to moderate differences in criterion dispersion before logarithmic dispersion is used to derive the final criterion weights. The proposed method is evaluated using a supplier selection dataset consisting of nine alternatives and eight criteria and is integrated with four MCDM methods, namely SAW, MOORA, WP, and WASPAS. The results show that LOGSTA-D produces a more moderated distribution of criterion weights by reducing the dominance of highly dispersed criteria and redistributing importance toward criteria with lower initial weights. Across the evaluated MCDM methods, the resulting rankings exhibit a high degree of consistency, with limited changes among intermediate alternatives. The robustness analysis shows that the LOGSTA-D with MCDM framework achieves a robustness value of 0.9875, compared with 0.9604 for the LOGSTA with MCDM framework, indicating a higher level of ranking consistency under the evaluated conditions. These findings demonstrate that LOGSTA-D provides a controlled dispersion-based weighting mechanism that can support robust decision-making in supplier selection problems.]]>
Copyrights © 2027
