MCDM'07 - paper no. 10

 

Back to MCDM'07 contents
 

Reference point method with lexicographic min-ordering of individual achievements

Włodzimierz Ogryczak

Abstract:

The reference point method (RPM) is a very effective technique for interactive analysis of the multiple criteria optimization problems. It provides the DM with a tool for an open analysis of the efficient frontier either connected or disconnected. The interactive analysis is navigated by the commonly accepted control parameters expressing reference levels for the individual objective functions. The individual achievement functions quantify the DM's satisfaction from the individual outcomes with respect to the given reference levels. The final scalarizing function is built as the augmented max-min aggregation of individual achievements which means that the worst individual achievement is essentially maximized but the optimization process is additionally regularized with the term representing the average achievement. The regularization by the average achievement is easily implementable but it may disturb the basic max-min aggregation. In order to avoid inconsistencies caused by the regularization, the max-min solution may be regularized according to the lexicographic min-order thus leading to the nucleolar RPM model. The nucleolar RPM implements a consequent max-min aggregation taking into account also the second-worst achievement, the third-worse and so on, thus resulting in much better modeling of the reference levels concept. The lexicographic min-ordering regularization is more complicated in implementation due to the requirement of point-wise ordering of partial achievements. Nevertheless, by taking advantage of piecewise linear expression of the cumulated ordered achievements, the nucleolar RPM can be formulated as a standard lexicographic optimization. Actually, in the case of concave piecewise linear partial achievement functions (typically used in the RPM), the resulting formulation extends the original constraints and criteria with simple linear inequalities thus allowing for a quite efficient implementation. It can be also approximated with the analytic form using the ordered weighted averages. The paper analyzes both the theoretical and practical issues of the nucleolar RPM.

Keywords:

Multicriteria optimization, reference point method, lex-min order

Reference index:

Włodzimierz Ogryczak, (2008), Reference point method with lexicographic min-ordering of individual achievements, Multiple Criteria Decision Making (3), pp. 155-174

Full text:

download

Scopus citations in 1 paper(s):
  1. Ogryczak, W., & Kozłowski, B. (2011). Reference point method with importance weighted ordered partial achievements. TOP, 19(2), 380-401. doi:10.1007/s11750-009-0121-4