Non-uniform allocation of information granularity to improve consistency and consensus in multi-criteria group decision-making: Application to building refurbishment

Abstract

If a decision depends on the views of a group of people, it is crucial that these views are consistent, in particular if they are represented in terms of fuzzy preference relations. Moreover, before making the decision, the agreement achieved must be as high as possible. In most cases, to achieve it, the initial viewpoint of each group member must be modified during the decision-making process, which causes an information loss. To handle this issue, some conditions have been added to the modification process. Concretely, some approaches have turned to an allocation of information granularity to control it. However, these approaches either allocate the information granularity in a uniform way or not consider the issues of consistency and consensus in a scenario of multi-criteria group decision-making. To enhance these approaches, we elaborate a novel granular-based approach composed of two processes. Firstly, an automatic process based upon a non-uniform allocation of information granularity is developed to maximize both the consensus and the consistency in multi-criteria group decision-making where fuzzy preference relations are used to model the group members’ views. Based upon it, secondly, an interactive process requiring the implication of the group members is introduced, which also intends to maximize both the consensus and the consistency. Some experimental studies are completed to show the essence of this new approach and to demonstrate its performance and flexibility. A case study on building refurbishment is also conducted to validate its effectiveness and feasibility in practical real-world problem-solving.