Modelling Self-Led Trust Value Management in Grid and Service Oriented Infrastructures: A Graph Theoretic Social Network Mediated Approach

Introduction

The main motivation of this work is to provide a trust-based model that brings together ideas from previous research, and consolidates these in a graph-theoretic framework. The model proposed is simple and computationaly lightweight meaning that it has practical applications. The model describes a self-led trust approach, giving a level of autonomy to trust calculations, not simply being dependant on previous scores from a trust chain/neighbourhood, leading to more accurate trust levels. The trust scores should also be mediated, not just dependant on one trust score but allowing some level of control by enabling individuals to weight the available information in favour of more trusted sources and to better reflect the individuals own trust strategies. The extant trust related research literature is vast and highly diverse. Trust has been studied from each and every angle: in the philosophical, sociological, psychological, computer scientific, economic, and legal sense – just to name a few (Karvonen, 2007). One perennial barrier to synthesising a definitive e-service trust model and theory of trust from this work is the lack of agreement as to definitions of trust (Grabner-Krauter et al., 2005; Grabner-Krauter & Kaluschia, 2003). Thus, inadequate trust conceptualisations and a lack of a unifying viewpoint or paradigm have frequently lead to weak theoretical rationale for empirical studies and a consequent inability to develop coherent and efficient theories. Indeed, one of the central difficulties is that the notion of trust is closely related to other concepts such as reliance, competence, trustworthiness and credibility. Nevertheless, within the vast extant trust literature it is possible to identify a core body of classic work. Further, we have chosen to focus on the notion that trust as a definite measurable confidence level, mediated by inherent risk. Classic trust studies have firmly established the notion that trust and distrust are threshold points on a continuum of probability assessment (Gambetta, 1988). Thus, we trust an entity to perform a particular task if and only if the likelihood that the entity will fulfil its obligation lies above a particular threshold value. This notion is central to the consumer and provider e-service model that is presented later in this paper. A concise review of the relevant literature now follows so as to partially contextualise the model.