This paper addresses the problem of composing different ambiguous and vague pieces of information without a central authority through a set of distributed agents, each with limited perspective, converging to a shared point of view by exchanging information only with a reduced subset of nodes (i.e., their respective neighbourhood). To this end, the distributed consensus problem is extended in the fuzzy fashion. As a result, the framework allows to compose several heterogeneous and ambiguous/linguistically expressed opinions in a decentralised way, both in terms of value with higher belief and in terms of ambiguity associated to the final agreement value. The proposed framework is applied to a case study related to crisis management for critical infrastructures, where human operators, each able to observe directly the state of a given infrastructure (or of a given area considering the vast and geographically dispersed infrastructures), reach a distributed consensus on the overall criticality of a situation expressed in a linguistic, fuzzy way. Such a consensus is reached in terms of actual severity of the scenario (single integrators) or in terms of both severity and evolution tendency (double integrators).
Keywords: critical infrastructures, fuzzy systems, consensus, distributed systems