A linear programming model dealing with ordinal ratings in policy capturing of performance appraisal

Linear programming (LP) has been utilized as an alternative to policy capturing for inferring criteria weights in a performance appraisal. Previous studies for policy capturing LP focused only on the criteria of numerical ratings while excluding ordinal ratings. This study presents an extended LP model to deal with ordinal ratings. Since LP can handle only numerical ratings, a set of possible ratings for each ordinal criterion is transformed into a set of unknown numerical ratings by introducing a function based on elements' relations of ordinal ratings. Then, the unknown numerical ratings are connected to a ‘weak relation’ or ‘strong relation’. Solving our model gives the optimal values for the criteria weights and the unknown numerical ratings. A simple example is given to illustrate our model.