BUILDING REPUTATION FOR SERVICE-ORIENTED AMI: MODELING, ALGORITHMS, AND ANALYSIS

Service reputation is a key factor for service selection and service composition in
Service-Oriented Ambient Intelligence systems. Hence, service reputation computing should fully
reflect the feature of multi-rating fusion and the utility value dynamic attenuation characters of the
rating. The paper combines D-S evidence theory with dynamic attenuation and puts forward a
service reputation computing algorithm based on multi-rating fusion, which is adapted to the
Ambient Intelligence systems. First, a layered computing model of the service reputation is given.
Then, a mechanism of dynamic attenuation based on time windows, an objective rating and
advertisement honesty rating of service, and a user credibility computing algorithm are presented.
Afterward, the rating information is combined with the D-S evidence theory to raise an
aggregation algorithm of the service general reputation for the Ambient Intelligence environments.
Finally, a prototype test is carried out to verify the effectiveness and availability of the model
together with the algorithms.