A Study of Highways Performance with a Molp-Dea Model and an Interactive Tri-Criteria Linear Programming Package (Trimap)

This paper deals with the evaluation of a few Brazilian privatised highways. We use Data Envelopment Analysis (DEA) for that. As the number of highways is too small, we need to use an additional method to avoid ties in the evaluation index. The MCDEA model of Li and Reeves (a tri-objective model) is adequate to solve this problem. The authors suggest the extension of the CCR model, considering two extra objective functions, i.e. a min-max function, corresponding to a measure of equity and a min-sum function, favouring some sort of overall evaluation of the DMUs, from the point of view of the DMU under analysis. The graphic decomposition of the weighting space (used in TRIMAP as an operational tool), the possibility of introducing constraints on the objective function values during the search automatically translated into the weighting space make it specially adequate to explore the MCDEA model.