An interpolation-based method for the time weighed vector elicitation in temporal PROMETHEE II applications

Abstract

Many real-life decision problems are time-dependant; they cannot be effectively addressed with the classic multi-criteria decision aid methodology which deals with 'static' problems. In fact, comparing time-dependant alternatives (possible decisions, actions …) may require several assessments spaced in time. If the aim is to rank these alternatives by aggregating their periodic assessments, one would wonder how to weigh these assessments with respect to the time scale. This paper introduces a novel method -based on interpolation- for the elicitation of the periodic weights required in temporal PROMETHEE II. Assuming that each periodic assessment has a relative temporal weight, the proposed method derives, from the decision maker, a subset of these weights. Then the remained ones are found by an appropriate linear interpolation. Simulation results show that a few elicited weights are sufficient to determine an effective approximation of the whole set of the desired temporal weights.