A Framework for Automatic Generation of Processes for Self-Adaptive Software Systems

The self-adaptation of software systems is a complex process that depends on several factors that can change during the system operational lifetime. Hence, it is necessary to define mechanisms for providing a self-adaptive system the capability of generating during run-time the process that controls its adaptation. This paper presents a framework for the automatic generation of processes for self-adaptive software systems based on the use of workflows, model-based and artificial intelligence planning techniques. Our approach can be applied to different application domains, improves the scalability associated with the generation of adaptation plans, and enables the usage of different planning techniques. For evaluating the approach, we have developed a prototype for generating during run-time the workflows that coordinate the architectural reconfiguration of a web-based application.