A reputation scheme with witness reasoning for service selection

Service consumers would have different degrees of satisfaction on a service and a collection of these impressions over a period of time within a network give service reputation. Service reputation can be propagated via referral network recommendation. Reputation can be a criterion for service selection. The proposed service selection model is able to estimate Web service reputation by collecting customer witnesses to form fuzzy inference rules and to solve the conflicts via compromise strategy. It used fuzzy Petri net (FPN) to simulate referral network and outrank services based on their reputations. Finally, an example of reputation-based Web services selection is used to demonstrate the proposed approach.