Enabling Adaptation in Trust Computations

Abstract

Abstract-Digital systems have been rapidly evolving within highly dynamic and unstructured environments, where the lack of a central authority forces entities to interact with each other through collaboration and negotiation. Digital agents often use Trust models in order to compute the level of trustworthiness of the partner they want to collaborate with. Unfortunately, due to the evolution speed of open and collaborative environments, the trustworthiness of an agent varies over time, and as a result, Trust models must be continuously adapted to the changing context. In this work we address the problem by presenting a self-adaptive model for Trust computations. In particular, the proposed methodology seeks to continuously align the trust model in force with the changing context in Web 2.0 dynamic applications such as forums, blogs, p2p systems. The self-adaptation is reflected in the auto-organisation of the Trust function to obtain an accurate degree of agents' trustworthiness.