A multi-agent assisted approach for spatial Group Decision Support Systems: A case study of disaster management practice

Abstract

The paper proposes an alternative new approach in contrast with the traditional methods to deal with multi-criteria group decision-making problems. It takes into account the multi-criteria group decision-making process as a multi-stakeholder multi-issue negotiation problem, in which stakeholders attempt to lead a consensus on the relative importance of the criteria by using software agents. To do so, it suggests three main steps: pre-negotiation, automated negotiation, and evaluation phases. The pre-negotiation phase is a human-computer interaction by which software agents attempt to exhibit and model the preferences space of the stakeholders. In the automated negotiation phase, the agents come together to negotiate on the criteria weights to reach an agreement on behalf of the stakeholders. Finally, in the evaluation phase, the evaluator agent applies a sensitivity analysis method to determine output variations due to the inputs and parameters.The proposed method is applied to a disaster management practice as a real-world case study, in which some stakeholders jointly attempt to identify the strategic roads in disaster situations specifically, flood events. Three spatial criteria are used for evaluating the road transportation network: load capacity, access to emergency suppliers, and importance of the roads in geometric structure of the network. The results of the study confirm that the proposed method is an efficient alternative approach to deal with multi-criteria group decision-making problems.