Information Aggregation Approaches for Multi-criteria Applications

Abstract

The effective use of multi-criteria techniques in specific applications will depend on relevant available data. Often such data arises from several various sources, and it is then typically desirable to aggregate some of this data before use as this could reduce computational complexity. Additionally such data frequently involves uncertainty, and this must be considered in the aggregation. Aggregation approaches ranging from possibilistic conditioning to probability and possibility transformations can be used to produce homogenous representations (probability or possibility) for aggregations. Here we review several approaches to uncertainty aggregation that apply to multi-criteria systems. Additionally, we consider approaches to evaluate the uncertainty of the aggregation to provide guidance of the use of the aggregated data for multi-criteria applications.