Multidimensional Context Representations for Situational Trust

Abstract

We propose a generic model for situation representation in agent trust models. As we argue that general trust models are inappropriate in complex environments and their use for advanced decision-making is limited due to the lack of detail, we introduce a generic framework for context representation in trust model. Presented method is independent of the trustfulness representation formalism in the underlying trust model. It doesn't prescribe any particular form for the set of reference contexts that holds the trustworthiness information in the context space: we define and evaluate the baseline approach with regular grid and discuss future extension towards more adaptive approaches using octant trees and fuzzy clustering techniques