A method of multiobjective decision making using a vector value function

Abstract

A decision situation with partial information on preferences by means of a vector value function is assumed. The concept of minimum value dispersion solution as a reference point joined with a pseudodistance function from such a point and a dispersion level ε, lead to the notion of ε-dispersion set. The dispersion level represents the amount of value that the decision maker can be indifferent to, therefore he should choose his most preferred solution in this set. Convergence properties, as well as an interactive method based on the reduction of ε-sets by means of parametric variation of ε, to aid decision making in discrete problems is considered. Detailed numerical examples are included.