A METHOD TO SIMPLIFY PATTERNS WITH PROBABILISTIC STRUCTURE IN WEB SERVICE COMPOSITION

Abstract

One of the challenging issues in the field of web services is web service composition. A hybrid model of web services is the interaction between a number of web services for a specific purpose. For each web service, there are candidate services with the same function but different quality parameters. So far, many methods have been presented for the optimal web service composition, but rarely a method has been presented for web services composition with complex probabilistic structures. In this article, we first identified two new patterns of nested and parallel loops. Then, in order to help the optimal candidate selection, a new method was proposed to simplify the web services composition. In this method, the web-service composition graph is divided into two ways: 1) web-service composition graph without a conditional pattern and 2) web-service composition graph with the conditional pattern. In graphs with conditional patterns, the probability of each path is converted to the number of repetitions of that path, and then the NSGAII algorithm is used to determine the best path in the graph and find more realistic solutions, which on average leads to improvement of quality parameters of availability by 17 percent, response time by 16 percent and cost by 0.07 percent compared to the probabilistic simplification method.