Development of the PSI-RAPS-MCRAT method: a hybrid model

This paper presents a study on the MCDM (multiple criteria decision making) hybridization methods. Three MCDM methods are used in the hybridization process: PSI (Preference Selection Index), RAPS (Ranking Alternatives by Perimeter Similarity), and MCRAT (Multiple Criteria Ranking by Alternative Trace). PSI has long been known as an MCDM method, but it has an unexplored utility: the ability to self-calculate the weights for the criteria. Because of this, PSI is also known as an MCDM method that does not require users to calculate the weights for the criteria. RAPS and MCRAT are two MCDM methods that require users to calculate the weights for the criteria. The hybridization of PSI with RAPS and MCRAT aims to use the criteria weights calculated by the PSI method to serve the ranking of alternatives using the RAPS and MCRAT methods. This hybrid model of the three methods is named the PSI-RAPS-MCRAT hybrid model. Thus, applying the PSI-RAPS-MCRAT hybrid model to rank alternatives will generate three sets of ranking results: one from using PSI, one from using PSI-RAPS, and one from using PSI-MCRAT. The accuracy of the PSI-RAPS-MCRAT hybrid model was tested through a specific problem of ranking welding robot types. The number of welding robots to be ranked is seven, while the number of criteria to reflect each type of robot is six. Spearman's rank correlation coefficient was used to evaluate the accuracy of the proposed model. The lowest being between the PSI method and the PSI-RAPS method at 0.8214, and the highest being between the PSI-RAPS method and the PSI-MCRAT method at 0.9643. The results show that the PSI-RAPS-MCRAT hybrid model has high accuracy. This study has provided a completely new and accurate method in the field of MCDM