D-STEM: a modified step method with desirability function concept

Abstract

Step method (STEM) is one of the well-known multi-objective optimization techniques. STEM has proven to be effective in extracting a decision maker (DM)'s preference information for a satisfactory compromise. However, it has been criticized for not considering the differing degrees of satisfaction associated with an objective function value, and for not providing flexible options in the process of preference information extraction. This paper proposes a modified STEM, called D-STEM, to overcome the methodological limitations of STEM. D-STEM utilizes the concept of a desirability function to realistically model the differing degrees of satisfaction. D-STEM also allows a DM to choose either tightening or relaxation, which makes the preference articulation process more efficient and effective. The advantages of D-STEM are demonstrated through an illustrative example. (c) 2004 Elsevier Ltd. All rights reserved.