Partial Information and Sensitivity Analysis in Multi-objective Decision Making

Many decision aids have been designed to help decision makers choose between several alternatives to deal with some or all of the problems arising from conflicting objectives, uncertainty, groups of decision makers, etc.

Abstract

Many decision aids have been designed to help decision makers choose between several alternatives. Such aids are typically intended to deal with some or all of the problems arising from conflicting objectives, uncertainty, groups of decision makers, etc. In all quantitative analysis is essential to check the sensitivity of the conclusions to the data. In designing decision aids this is even more true as the decision makers become aware of the implications and possible inconsistencies of their judgements. Sensitivity analysis can focus on those judgemental inputs which are more important and, therefore, need to be revised the most carefully. In this paper we review some of the previous approaches to sensitiviy analysis. With a few exceptions, they consider sensitivity to one or, at most, two data inputs at a time, the remaining data being taken as fixed. With recent advances in optimisation theory and with the advent of cheap computer power, it seems possible to provide a more general approach to sensitivity analysis, allowing for simultaneous variation in all the data. We briefly introduce some tools to help in this task.