Modelling uncertainty in performance appraisal process based on evidence theory and fuzzy theory

Uncertainty involved in human judgments is an important cause of error and loss of credibility in outputs provided by performance appraisal (PA) processes. In few existing studies related to the errors and uncertainties in PA process, solutions mainly based on the use of fuzzy tools have been presented in this regard. These solutions have fundamental deficiencies such as inability to cope with epistemic uncertainty and also problems associated with their implementation. Considering these problems, in this paper, a new model based on evidence theory and fuzzy tools has been proposed to model uncertainty in PA process. The proposed model makes it possible for assessors to provide their ratings commensurate with their level of knowledge and also has the ability to deal with uncertainty caused by randomness and ignorance. In the proposed model which has been designed based on multi-source assessment framework, the uncertainty contained in the data obtained from two common evaluation scales, including Visual Analogue Scale and fuzzy linguistic scale, along with data related to the reliability of evaluator, have been modeled in evidence theory structure. These data then have been aggregated with evidence combination rules. The performance, benefits and improvements resulting from the proposed model, compared with other common aggregation methods in P.A models, have been investigated using simulated data and a numerical example. The results show that the proposed model in addition to improving the ability of dealing with uncertainty in P.A processes and facilitating announcing opinion by raters, provides more accurate results than traditional P.A models.