On characterizing features of OWA aggregation operators

We introduce the ordered weighted averaging (OWA) operator and emphasize how the choice of the weights, the weighting vector, allows us to implement different types of aggregation. We describe two important characterizing features associated with OWA weights. The first of these is the attitudinal character and the second is measure of dispersion. We discuss some methods for generating the weights and the role that these characterizing features can play in the determination of the OWA weights. We note that while in many cases these two features can help provide a clear distinction between different types of OWA operators there are some important cases in which these two characterizing features do not distinguish between OWA aggregations. In an attempt to address this we introduce a third characterizing feature associated with an OWA aggregation called the focus. We look at the calculation of this feature in a number of different situations.