Granada 13 - 14 February, 2014
The notion of consensus plays a key role in modelling group decisions, and for a long time it was meant as a strict and unanimous agreement, however, since various decision makers have different more or less conflicting opinions the traditional strict meaning of consensus is unrealistic. The human perception of consensus is much “softer”, and people are willing to accept that a consensus has been reached when most or the more predominant actors agree on the preferences associated to the most relevant alternatives. The “soft” meaning of consensus, advocated as realistic and humanly consistent, can lead to solve in a more constructive way group decision making situations by using modeling tools based on fuzzy logic.
In this presentation, the problem of modeling consensus under individual fuzzy preferences is considered, and two different models are synthesized. The first one is static and is based on the algebraic aggregation of the individual preferences aiming to find a consensus defined as the degree to which most of the important individuals agree as to their preferences concerning almost all of the relevant alternatives. The second one is dynamic and it combines a soft measure of collective disagreement with an inertial mechanism of opinion changing aversion. It acts on the network of single preference structures by a combination of a collective process of diffusion and an individual mechanism of inertia.
The first application is focused on consensus evaluation by computing the importance of decision makers in terms of their influence strength in a social network, starting from centrality measure and combining it with the fuzzy m-ary adjacency approach. Accordingly, a flexible consensus measure is introduced taking into account the influence strength of the decision makers according to their eigenvector centrality.
The second applications addresses the problem of representing fraudulent attacks using attack trees as a component of a multi-agent system. Assuming that the opinions of experts involved in the design of the attack tree are represented by fuzzy preference relations, a dynamical consensus model is introducing aiming at finding a shared representation of the attack tree.