Opsearch, Vol. 35, No. 2, 1998
0030-3887/95 $ 2.00 + 0.00
© Operational Research Society of India
Solving Discrete Multiobjective Optimization Problems
Based on Approximation
Alfonso Mateos, Sixto Rios-Insua and Gregorio Nevado
Department of Artifical Intelligence, School of Computer Science
Technical University of Madrid, Campus de Montegancedo
Boadilla dei Monte, 28660 Madrid, Spain
Abstract
We consider the descrete multiobjective decision making problem under partial
information given through a vector value function. We propose adecision aid
procedure based on an approximation to the efficient set and its interactive
reduction. Finally, an example iIIustrates the method.
1. INTRODUCTION
In multiobjective decision making the set of efficient solutions plays an
important role in the solution process. Its main property is that from a sulution
in this set, it is not possible to move feasible so as to increase one of the
objectives without necessarily decreasing at least one of the others. However,
the generation of such set may be involved in linear problems and very difficult
in nonlinear ones. In the case of discrete problems this is easier, but we note
that the set of efficient solutions might be far too be for practical purposes.
Hence, the generation of the efficient set is not usually considered the resolution
of a multiobjective problem because this set might have many elements and is
not generally totally ordered. Thus, there is a need for more clever strategies to
generate a set of representative efficient solutions or an approximation that gives
a fair representation of the whole efficient set. Although there are several methods
to aid adecision maker (DM) to generate the efficient set or a representative
subset (see, e.g., [1], [5]. [6]. [11]. [14]. [15], [16]. [17]. [18].) under different
approaches, there is not a definitive solution to this problem.
We consider in this paper a value-based procedure (e.g., [2]. [3], [7]) with
partial information on the DM's preferences, which uses an approximation concept,
[8], [9] and intends, on one hand, to make easier the generation process of a
representative set of the whole efficient set and, on the other its interactive
reduction to reach a final solution.
Throughout the paper we employ the following notation: For two scalars a
and b, a b denotes a > b or a = b. For two vectors