Abstract:

In order to tackle the problem appeared in dynamic and stochastic nature of web service, a new web service selection algorithm based on interval quality of service (QoS) is proposed. The uncertainty is described by interval QoS with probability distribution. The probability distribution can be computed in terms of the sample data. By comparing the probability density in the confidence interval of QoS, it can more accurately describe the degree with which the QoS is affected by the open environment. Integrating this comparison method with interval QoS model of composite web service, a web service composition algorithm with the interval QoS guaranteed is designed. Experiment shows the interval numbers contribute to reduce the effect imposed by the dynamic environment and enhance the correctness of service selection.

Key words: web service selection, uncertainty, quality of service, interval number