Novel Weight Stability Analysis for Net-Flow Based Multiple-Criteria Method Applied to Cources of Action Analysis


  1. Guitouni, A.
  2. Lang, P.
Corporate Authors
Defence R&D Canada - Valcartier, Valcartier QUE (CAN)
This report presents a novel type of approach for weight stability analysis developed for multiple criteria methods, which is implemented in the Commander’s Advisory System for Airspace Protection (CASAP) prototype. Criteria weighting is a complex preference elicitation process. In these circumstances, it is helpful to determine to what extent the solution (ranking) obtained is sensitive to variations in the relative importance coefficients. Stability analysis provides the decision maker (DM) with a precise idea about the sensitivity of a decision to any change in the weighting parameters. In this report we developed, implemented and tested a novel type of stability analysis approach for Multiple Criteria Decision Analysis (MCDA) net flow based methods. This approach is based on the concept of searching for a maximal geometrical box around a centric point (centroid). Then, a multiobjective problem is formulated to find the maximal stability zones around the weight vector fixed by the DM. The problem is then transformed into a lexicographical mathematical program. The implementation of the approach and some experimental results are presented. The empirical results are very significant: It is possible to identify the most sensitive criteria in the decision outcome. It is also possible to identify the most conflicting ones. This report also discusses how we think such an approach could be extended to other MCDA methods as well as how it could be used for machine learning. For

Il y a un résumé en français ici.

Report Number
DRDC-VALCARTIER-TR-2001-216 — Technical Report
Date of publication
01 Apr 2008
Number of Pages
Hardcopy;CD ROM

Permanent link

Document 1 of 1

Date modified: