Research on the Improvement of Traditional Linear Weighted Algorithm for QoS-based Web Service Selection

Abstract

At present, quality of service (QoS) has become an important guidance in selecting a satisfactory web service when many similar web services provide overlapping or identical functionalities. Therefore people have carried out many studies on definition and standardization of QoS properties and QoS-based selecting algorithms of web service. Current web service selection usually adopt linear weighted algorithm. But it still has some limitations. This paper proposes a new QoS-based algorithm developed from traditional linear weighted algorithm for web service selection. The main improvements include removing distinction of QoS properties value tendency, selecting fixed value and a hybrid subjective-objective approach to set the weight of each QoS property. We also present the analysis of the time complexity. Finally we do a comparison experiment between our algorithm and traditional algorithm. The results show that the improvements are feasible and our algorithm has better performance.