Decision Supporting Approach under Uncertainty for Feature-Oriented Adaptive System

A self-adaptive system is able to adjust its behavior in response to its perception of the environment. One challenging issue for a self-adaptive system is to deal with its variability. The self-adaptive system should manage their diverse functionalities and behaviors and select one of them depending on the encountered environments. To handle this variability, many adaptive systems adopt the feature model to represent their architecture. A feature model is able to set up all possible system architectures by representing the available alternatives and their relations. However, the relations between features are complex and intertwined with each others. These characteristics cause uncertainty which disturbs the decision making. Uncertainty means the incomplete and insufficient condition to make a decision. Thus, determining an appropriate feature under uncertainty is still a challenging problem. In this paper, we propose a modeling process to support decision making under uncertainty for feature-oriented adaptive systems. By following the proposed process, the engineer is able to extract and design decision support model from the system feature model. In the decision support model, to mitigate uncertainty, we fragmentize the system features into several pieces including only highly relevant features with probability values between intra-and inter-fragments. Through these probabilities, the system analyzes the impact of uncertainty on a decisions and determines an appropriate decision with this analyzed information.