Considering Bipolarity of Attributes With Regards to Objectives in Decisions Evaluation

Abstract

Many decision problems in engineering, economics, and management consist of selecting and/or ranking alternatives that are characterized by multiple attributes in order to satisfy multiple and possible antagonist objectives. Besides this double multiplicity aspect, decision making mainly in engineering economics (decision problems where potential solutions of a problem must be economically viable along with some technical aspects) generally necessitate to cope with many stakeholders opinion regarding the importance to assign to each attribute and/or each objective as well as uncertainty that can impact any of these components (attributes values, objectives measurement and stakeholders statements). Decision processes in this context go through some steps such as 1) definition of the overall decision goal (find the best sustainable site to build a waste management facility for instance), this goal can further be split into many objectives to satisfy; 2) identification of potential alternatives; 3) elicitation and evaluation of attributes that characterize alternatives; 4) and finally evaluate alternatives to recommend the “good enough” alternative to implement. The aim of the research undertaken in this paper is to address the question raised in the third and fourth points by exploiting the bipolar nature of attributes with regards to objective to elicit and assess attributes and to build a final recommendation procedure. Indeed, we will highlight the bipolar nature of attributes with regards to pursued objectives (given an objective, there will be attributes that act in the sense of realization of this objective, that we refer to as supporting attributes and those working against the achievement of this objective, referred to as rejecting attributes) to derive a framework for elicitation and assessment of attributes values. Finally, relying on the bipolar nature of obtained attributes, satisficing game approach is used as a tool to support the evaluation and final recommendation process. The approach is applied to a real world problem which aim is to select the best sustainable site to build a waste management facility.