A web service composition framework in a heterogeneous environment
Naoufal El Allali,Mourad Fariss,Hakima Asaidi,Mohamed Bellouki
DOI: https://doi.org/10.1007/s12652-022-03761-9
IF: 3.662
2022-02-25
Journal of Ambient Intelligence and Humanized Computing
Abstract:With the growing number of web services and the diversity of customers' requirements on the Web, it is still a challenge to develop a specific service in a heterogeneous environment dynamically. Given that most web service providers cannot always provide a service that meets a complex request, the service composition mechanism is required to resolve this problem. However, it constitutes a real challenge to create a dynamic and self-adaptive web service composition system to respond in real-time to a large scale of services. Furthermore, several criteria affect the composition process, including the number of services, the size of the graph, the run-path, the execution time, and the parallel execution. In this paper, we propose a new framework to efficiently solve the composition problem based on the cooperation of the mobile agents and an improved Ant Colony Optimization (ACO) algorithm. Firstly, a complex request is transformed into sub-problems to reach the potential sub-solutions. Secondly, the ACO is applied to find the optimal sub-solutions regarding the number of services and the run paths. In parallel, a crucial mobile agent named FRIEND-AGENT aims to reduce impractical paths and overlapped solutions during the solution building process by artificial ants. This cooperation leads to achieve an optimal service composition with less memory consumption, an optimal graph, and improved scalability of our proposed system in real-time. The proposed framework has been validated as high performance with eight different public repositories due to the high efficiency to resolve the composition problem in all the scenarios. Moreover, it has been proved that our approach excels the state-of-the-art approaches to find the optimal solutions concerning the execution time, the number of services, and the execution path.
computer science, information systems,telecommunications, artificial intelligence