Engaging the Dynamics of Trust in Computational Trust and Reputation Systems

CRM: An efficient trust and reputation model for agent computing

“…The reason behind this simplification is that our main contribution in this paper is to equip the agents to efficiently evaluate the trust and get adapted with continuous environment changes. Although the domain is important in trust evaluation (as mainly considered in some trust-based frameworks [26,7]), in this paper we only focus on the adaptation of agents with dynamically changing environment and on how agile the agent is while acting where the trust evaluation is crucial. Furthermore, although the time is omitted from the formulation, it is implicitly represented as the trust function is continuous over T .…”

“…Ag a can be computed by Equation 1. This method of direct trust evaluation is similar to the ones proposed in [7,14]. In this Equation, the ratio of the "number of successful outcomes" to the "total number of possible outcomes" is computed, where w i is the weight associated to the interaction type i to represent its importance and v ij is the value of the interaction, which is particularly important in transactional settings to avoid two transactions with different values being treated equally.…”

“…There are some similar approaches in the literature addressing this issue. For example, in [7], the authors discuss the limitations that are used in the freshness of the data to be evaluated. In our model, we assess this factor denoted by T iR(∆t Ag b Aga ) ij by using the function defined in Equation 2 and we do not make the system so sensitive to the past data as it might bring up more confusion to the trustor agent.…”

“…In this section, we analyze some recently emerged systems like TRSIM [6], Repage [21], Formal [30], and HIT [29]. We also analyze and compare some related robust frameworks like [7,26,14]. These models are worthy to be discussed in terms of their attempt to address the trust establishment in a different perspective.…”

“…In our model, the issue is managed by considering the witnesses trust and our merging method takes into account the proportional relevance of each reputation value, rather than treating these values equally. In [7], a comparison between two trust approaches is discussed. The first approach computes the weighted trust by aggregating the past information (history) with respect to the freshness.…”