Evaluating Web Service QoS: A neural fuzzy approach

In Service-Based Applications (SBA), evaluating Quality of Service (QoS) values is an important issue and difficult to calculate due to uncertain and poor available information about quality parameters as well as non-linear relation between them. Fuzzy Logic (FL) can provide such a relation by employing if-then rules using inference methods. However, in a Fuzzy Inference System (FIS), fuzzy rules and membership functions are fixed and can only be defined by an expert of the system. In most cases, there is a need for an adaptive model to dynamically obtain fuzzy rules, optimise membership functions, and predict the state of QoS in the system. Therefore, we propose a fuzzy adaptive system based on learning techniques that can adjust the membership functions initially defined by experts. In particular, we use Adaptive Neuro Fuzzy Inference System (ANFIS) to tune QoS parameters adaptively according to a given dataset. The difference between predicted and desired QoS is minimized during the learning process. The trained FIS can be used in decision making process for QoS-aware adaptation. Experimental results demonstrate the ability of the proposed approach in finding relationship between the quality variables and predicting the overall QoS values.