Journal of Systems Engineering and Electronics Vol. 21, No. 6, December 2010, pp.975–980
Available online at www.jseepub.com
Dynamic programming methodology for multi-criteria group
decision-making under ordinal preferences
Wu Li
1,*
, Guanqi Guo
1
, Chaoyuan Yue
2
, and Yong Zhao
2
1. School of Information and Communication Engineering, Hunan Institute of Science and Technology, Yueyang 414006, P. R. China;
2. Institute of Systems Engineering, Huazhong University of Science and Technology, Wuhan 430074, P. R. China
Abstract:
A method of minimizing rankings inconsistency is pro-
posed for a decision-making problem with rankings of alter natives
given by multiple decision makers according to multiple criteria.
For each criteria, at first, the total inconsistency between the rank-
ings of all alternatives for the group and the ones for every decision
maker is defined after the decision maker weights in respect to the
criteria are considered. Similarly, the total inconsistency between
their final rankings for the group and the ones under every criteria
is determined after the criteria weights are taken into account.
Then two nonlinear integer programming models minimizing re-
spectively the two total inconsistencies above are developed and
then transformed to two dynamic programming models to obtain
separately the rankings of all alternatives for the group with re-
spect to each criteria and their final rankings. A supplier selection
case illustrated the proposed method, and some discussions on
the results verified its effectiveness. This work develops a new
measurement of ordinal preferences’ inconsistency in multi-criteria
group decision-making (MCGDM) and extends the cook-seiford
social selection function to MCGDM considering weights of criteria
and decision makers and can obtain unique ranking result.
Keyw ords: multi-criteria group decision-making (MCGDM), ordi-
nal preference, minimum deviation method, dynamic programming,
Cook-Seiford social selection function.
DOI: 10.3969/j.issn.1004-4132.2010.06.008
1. Introduction
Decision-making is the procedure to find the best or most
suitable alternatives among a set of feasible alternatives.
In most organizations, group problem solving is an impor-
tant aspect of strategic decisions, for example on supplier
selection issues in supply chain management. Decision-
making in organizations requires the participation of man-
agers from a wide range of departments, with different
Manuscript received April 14, 2009.
*Corresponding author.
This work was supported by the National Natural Science Founda-
tion of China (60904059; 60975049), the Philosophy and Social Science
Foundation of Hunan Province (2010YBA104), and the National High
Technology Research and Development Program of China (863 Program)
(2009AA04Z107).
interests, backgrounds, experience, culture, etc. This di-
versity, as well as the complexity common to m any group
interactions, makes group decision-making burdensome.
Some important decisions are made by a group of deci-
sion makers according to multiple criteria, an d they are
called multi-criteria group decision-making (MCGDM)
problems. For example, multiple criteria including qual-
ity, delivery, performance history, warranties, price, tech-
nical capability and financial position need to be taken into
account simultaneously when selecting suppliers in supply
chains [1,2]. MCGDM often requires the decision mak-
ers to provide qualitative/quantitative assessments for de-
termining the performance of each alternative with respect
to each criterion, and the relative importance o f evaluation
criteria with respect to the overall objective of the prob-
lems. These problems will usually result in uncertain, im-
precise, indefinite and subjective data being present, which
makes the decision-making process complex and challeng-
ing [3]. Thus, decision makes usually are willing or easy
to give his ordinal preferences of alternatives with respect
to each criteria.
Some methods were proposed in literature published for
the decision-making with ordinal preferences of alterna-
tives [4–8]. However, all these methods were developed
for aggregating individual preferences of alternatives in
single criteria setting, or for combining the rankings of
the alternatives in respect to each criteria in single-person
multi-criteria decision. Two methods combining of TOP-
SIS and 0–1 programming is proposed for MCGDM with
the rankings of the a lternatives given by multiple deci-
sion makers according to every criteria [2]. But they may
bring out several different results for the same decision-
making problem. Reference [9] also generalized TOPSIS
to MCGDM with ordinal preferences and just considered
the overall weights o f decision makers. In fact, the author-
itativeness and validity of evaluation of alternatives given
by different decision makers in respect to the same criteria