Affiliations: Department of Computer Engineering, Yazd Branch, Islamic Azad University, Yazd, Iran
Correspondence:
[*]
Corresponding author: Sima Emadi, Department of Computer Engineering, Yazd Branch, Islamic Azad University, Yazd, Iran. E-mail: [email protected].
Abstract: The expansion of service-oriented architecture and the increasing number of web services has led to an increase in demand for their use. But since a single service alone may not be enough to meet the most relatively complex business processes requirements, it is necessary to combine several individual services to deliver user satisfaction. By increasing the number of services that have the same functionality, the quality of service provided by each service will play an important role in the service selection process; in the process of service composition, different services with different quality parameters come together to provide a new task. Therefore, offering the best quality service to the user is considered an important issue. The challenging issues in the service composition process include how to combine the web services with quality parameters based on user preference, long response time for the composition process, large search space, and the correlation between the services. In this paper, the quality-based service composition is modeled by considering the relationship between the services to improve the quality of service (QoS) parameters. The proposed model consists of several steps. In the first step, the inappropriate services will be pruned by applying the correlation between the services. In the second step, by determining the quality levels for the QoS, the APSO algorithm is used to select the best levels and, finally, the best services. In the service combination stage, the services selected from the previous stage are combined using a fuzzy genetic algorithm (FGA) to create a suitable combination service. The results show that when the correlation between the services is considered, the response time criterion improves significantly by integrating the quality parameters and pruning the candidate services, and reduces the search space.
Keywords: Context awareness, service correlation, genetic algorithm, fuzzy rules, service composition, particle swarm algorithm
