A modified common weight model for maximum discrimination in technology selection

摘要

This paper provides a modification of a recently developed common weight model for technology selection. The discrimination power of the existing model depends on a discriminating parameter. Although larger values of this parameter can give more discrimination, it should also be small enough to guarantee the existence of at least one efficient DMU. Hence, different values of this parameter should be examined to determine the proper value. In the model proposed here, the largest possible value of the parameter is utilised to maximise the discrimination power, and the existence of an efficient DMU is guaranteed by adding a few constraints. In addition, the paper presents comments on existing models, and the properties and the advantages of the model are explained. The contents of the paper are illustrated by several numerical examples.