Dynamic Architectural Selection: A Genetic Algorithm Based Approach

As the software industry is focusing on dealing with various requirements and environments, such as mobile and ubiquitous environments, software systems are increasingly undergoing many situational changes. These changes influence the quality of services that the software provides. Therefore, to maintain the performance of the software, it must be reconfigured. The reconfiguration is a complex problem if an application faces a large number of situations and has a number of software architectural instances. In this paper, we propose a novel approach to autonomous architectural selection in response to the current situation of various environments. This approach enables a software system to determine the best architectural instance for the current situation. To quickly find the best instance, we apply a genetic algorithm to the selection process. Further, we provide a performance evaluation to demonstrate that our approach efficiently find the best instance (or considerably good instance).