In this paper we address a ranking problem from a tender mechanism established by the Portuguese energy regulator to promote energy efficiency measures in electricity consumption. Measures candidates to this tender are subjected to previously known assessment rules, and weighted criteria. Our proposal has explicitly taken avoided consumption and cost of saved kWh as decision variables for defining two societal objectives. In order to define ranges for the weights of the criteria and reference performance values, the search for the most suitable sets of weights is done by an evolutionary algorithm. Measures were used that applied for funding in previous tender calls. The results demonstrate that in face of different sets of measures, the range of weights for each criterion may vary significantly. The proposal, besides helping to obtain a more suitable set of criteria weights, can also assist in the very process of formulating the criteria, in view of the regulator's societal objectives, since it was found that some of the used criteria actually lack the ability to discriminate between candidate measures. After a post-processing analysis, the DM is presented with a set of criteria weights from which he can select the one that better reflects his preferences. (C) 2012 Elsevier Ltd. All rights reserved.