A Composition of Web Services Using the Markov Decision Process and Long-Short Term Memory

The uncertainty of Quality of Service and the stochastic nature of web services make selecting an optimal web service composition a difficult task. The term “composition” refers to the process of combining the atomic web services into a complex task that satisfies all of the functionality of the requested service. The term “Quality of Service” refers to intangible features that are rarely the focus of academic inquiry. To standardize the skyline services, a relevance function is computed. For each service type, the top k web services are selected based on their relevance function, and then the Markov decision method is used to further refine the list of available services. Efficient web service selection and composition is maximized via the combination of a skyline, a Relevancy function, and a Markov decision process. The experimental findings support the improved performance of the proposed method.