Roghayeh hemmatjon - School of Electrical & Computer Engineering, University of Tehran
nasim nahavandi - Department of Industrial Engineering, Tarbiat Modares University
behzad moshiri - School of Electrical & Computer Engineering, University of Tehran

Recently, Coquet integral (CI) has been used in multi-criteria decision-analysis (MCDA) problems in which there are non-linear relations between decision criteria. The decision-maker imposes some information, on basis of preferences relations, on decision problem. The coefficients of evaluation method should be defined accurately in order to illustrate the intention of decision maker (DM). Different methods have been proposed in this area. The capacity definition methods based on optimization provide a solution space and pick the coefficients through it. In cases where there is a solution set rather than a unique solution, the coefficients are usually chosen arbitrarily from solution set, damaging the reliability of the result. In addition, solution usually provides more information compared to propositions made by DM. Therefore, DM may not fully interpret the result. So, robust capacity definition methods are proposed to overcome these drawbacks. On the other hand, these methods do not consider evenness (uniformity), which is a major property of capacity. Preferences of DM are made of a subset of alternatives called reference alternatives. When defined capacity sharply focuses on this information, it moves solution to the extreme points of solution space. Therefore, it loses its uniformity or as called evenness, and it is not reliable for evaluating non-reference alternatives. This paper uses an algorithm to define a capacity that is based only on the preference information of DM, and consequently, it becomes of DM’s preferences. Furthermore, it improves evenness of capacity and its reliability in evaluating non–reference alternatives. The algorithm is used to evaluate power plant projects which are huge national projects in Iran. Case–specific criteria are taken into account in addition to general criteria used in project performance evaluation. The evaluation results, obtained from the proposed algorithm, are compared with those of the most representative utility function method.

aggregation operators, non–additive robust ordinal regression, multiple criteria decision making, uniform capacity