Multiple attribute decision making with intervals for QoS-based web service selection

Traditionally asymmetric models are used for QoS specifications, average values are used for calculation, comparison or revision for attributes with indefinitive values (e.g., response time and reliability) under these models. These models are easy to program, but with negative affect on the accuracy of decisions. In this paper, a new interval based, symmetric QoS model is introduced, which is to achieve optimal solutions with improved TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) and multi-attribute decision making. Experiments show that by this approach, both the expression ability of the QoS model and the resolution of service selection are largely improved.