Robustness Analysis of an Outranking Model Parameters’ Elicitation Method in the Presence of Noisy Examples

Abstract

One of the main concerns in Multicriteria Decision Aid (MCDA) is robustness analysis. Some of the most important approaches to model decision maker preferences are based on fuzzy outranking models whose parameters (e.g., weights and veto thresholds) must be elicited. The so-called preference-disaggregation analysis (PDA) has been successfully carried out by means of metaheuristics, but this kind of works lacks a robustness analysis. Based on the above, the present research studies the robustness of a PDA metaheuristic method to estimate model parameters of an outranking-based relational system of preferences. The method is considered robust if the solutions obtained in the presence of noise can maintain the same performance in predicting preference judgments in a new reference set. The research shows experimental evidence that the PDA method keeps the same performance in situations with up to 10% of noise level, making it robust.

An Outranking Multicriteria Method for Nominal Classification Problems with Minimum Performance Profiles

“…According to [51], robustness is a key issue in the field of decision-aiding, as well as in operations research. As a result, numerous researchers have recently addressed this issue [51][52][53][54][55][56][57][58][59][60][61][62] and have proposed the use of performance measures for classification and clustering methods [63,64]. The term robustness refers to a capacity for withstanding "vague approximations" and/or "zones of ignorance" to maintain certain properties [51].…”

“…Thus, robustness with respect to different scenarios was assessed by changing some preference parameters, such as criteria weights, profiles of classes, and preference and indifference thresholds. As a result, a total of 138 scenarios were tested: the combination of changing the values of 13 criteria weights, four profiles of classes according to each criterion, and preference and indifference thresholds in ±10%, following similar procedures to those presented in [19,56].…”

Sovereign Credit Risk Assessment with Multiple Criteria Using an Outranking Method

“…To do so, a model that infers parameters is used by disaggregating preferences for a sorting method, proposed in [20]. Preference disaggregation consists of an indirect way of eliciting the decision maker's preferences, which arise out of preference examples [21]. In addition, a comparison is made between the results of the proposed model and the allocations of two risk rating agencies, Standard & Poor's and Moody's, for the year 2014.…”