Learning-based trust model for optimization of selecting Web services

Abstract

As the deployment of Web services increases in complex business application integration, it is becoming inevitable that several Web services may have the same or similar functionalities each holds different Quality of Services(QoS). In this context, it is becoming a challenge for Web services developers to select most pertinent services for their businesses. In this paper, we suggest a novel method so called Trust Model for selecting provider Web services among several similar providers but with different QoS. Our Trust Model is a function of historically gathered and learned QoS values, references or feedbacks from other services and honesty degree of this service. Client Web services satisfaction degree function was devised for checking the satisfaction degree when using our Trust Model. The proposed model is validated with initial result by comparing two cases: when using our Trust Model for selecting Web services and when not using our Trust Model. © Springer-Verlag Berlin Heidelberg 2007.