50 shades of partial information

Abstract

Since their origins in the 1980's, grey systems thinking and grey numbers have found a wide and growing range of applications. One of these is in providing decision support for multi-criteria choice. All the key elements of multi-criteria modelling are potentially infused with a lack of certainty: weights on criteria, performance levels of alternatives, choice of criteria, inter-dependence between criteria, and so on. For a long time, multi-criteria researchers have recognised that, especially from a practical point of view, their decision support models are imprecise, that they are often operating with only partial information about the problem they are seeking to solve. And, indeed, in some circumstances, especially in supporting choice between discrete alternatives, models do not need to be precise or based on full provision of information in order to be useful. This has been referred to as decision making with partial information and a wide variety of formalisations of multi-criteria decision making with partial information have been advanced. In this paper, we explore how using grey systems can provide a perspective on multi-criteria choice with partial information that potentially enhances existing thinking and directly relates to the ambiguity and uncertainty of choice - effectively the real world of decision modelling for many application contexts.