An integrated decision-making methodology based on Pythagorean fuzzy sets for social robot evaluation

The modern era robots are designed to realize many different tasks, and there are many types of them for different purposes. As robotic technology develops, new features are added to the robots, increasing their importance in human life. Being social is one of the preferred features of the new generation of robots. The socialness of the robots indicates the relationship between the robots and the humans. Thus, the robot evaluation process has great importance, including many different perspectives. Since the process contains a set of criteria and robot alternatives, it can be considered a multi-criteria decision-making (MCDM) problem. The evaluation process includes uncertainty and vagueness because of the humans' subjective evaluations. Pythagorean fuzzy sets (PFSs) can be effectively utilized as an extension of intuitionistic fuzzy sets (IFSs) to cope with uncertainty and vagueness in the decision-making processes. This study proposes an integrated fuzzy MCDM methodology based on interval-valued Pythagorean fuzzy sets (IVPFSs) for the social robot evaluation problem (REP). By assigning IVPFSs, it is aimed to represent uncertainty and hesitancy in an effective and flexible way by using both membership and non-membership functions. For this aim, a novel integrated methodology consisting of IVPF-based forms of DEMATEL, ANP, and TOPSIS methods is suggested. While DEMATEL and ANP methods based on IVPFSs have been, respectively, utilized to consider internal and external dependencies of criteria and calculate importance degrees of criteria, the social robot alternatives have been ranked via the IVPF-TOPSIS technique. These methods have been adopted with respect to decision makers' evaluations based on both linguistic variables and cardinal information. Since the high amount of data is in linguistic form, we also present a new way of fuzzifying the cardinal information for applicability. During a case study, five social robot alternatives have been evaluated in terms of five main and twenty-one sub-criteria. In order to present the effectiveness and the applicability of the proposed methodology, sensitivity and comparative analyses have also been conducted. A sensitivity analysis has been also applied to observe the shifts of the alternative ranks based on the changes of the criteria weights. Moreover, for verification of the obtained results IVPF-ARAS (Additive Ratio Assessment) method has been applied in the scope of comparative analysis.