A trust-aware framework for service selection and service quality review in e-business ecosystems

Publication Type:
Thesis
Issue Date:
2008
Full metadata record
Files in This Item:
Filename Description Size
Thumbnail01Front.pdf7.86 MB
Adobe PDF
Thumbnail02Whole.pdf162.58 MB
Adobe PDF
As e-Business has moved from a niche market to a decisive contributor for the success of most companies, some issues need to be solved in order to assist the continued success of e-Business. The challenge, to deploy fully autonomous business service agents which undertake transactions on behalf of their owners, often fails due to lack of trust in the agent and its decisions. Four aspects can overcome this challenge. Firstly, intelligent agents need to be equipped with self-adjusting reputation, trustworthiness and credibility evaluation mechanisms to assess the trustworthiness of potential counterparts prior to a business transaction. Secondly, such evaluation mechanisms must be transparent and easy to comprehend so agent owners develop trust in their agents’ decisions. Thirdly, the calculations of an agent must be highly customisable so that the agent owner can apply his personal experiences and security requirements to govern the decision making process of the intelligent agent. And finally, agents must communicate via standardised and open protocols in order to facilitate interaction between services deployed across different architectures and technologies. This thesis proposes the DEco Arch framework which integrates behavioural trust element relationships into various decision making processes found in e-Business ecosystems. We apply fuzzy-logic based soft computing techniques to increase user confidence and therefore enhance the adoption of the proposed assessment and review methodologies. A proof-of-concept implementation of the DEco Arch framework has been developed to showcase the proposed concepts in a case study and to conduct empirical experiments to evaluate the robustness and practicability of the proposed methodologies.
Please use this identifier to cite or link to this item: