An Approach to Controlling Context by Combining Game Theory and Control Theory for Self-Adaptive Software in Creative Computing

Abstract

Self-adaptive software is widely-use for users who need to output varies results after responding from changing environment. The results are normally followed by the Control Theory in order to be constructed in a rigid feedback loop. According to the Control Theory, the controller is designed to measure result errors for the purpose of re-inputs of edited results into the system. The research in the paper proposes an approach to improving the context-aware ability of controller by designing a controller that enhance the self-adaptive software to solve both certain and uncertain problems especially in business. The controller relies on both Control Theory based modules and combination based module, in order to generate outputs including but not limited to activities based on costs or balanced scorecards through adding business context sensors, and create feedback loop to test differences for each module. When the proposed controller is established in self-adaptive software, it is important to extract and filter business information due to user demands, and make software to react itself in a changing environment. In summary, a designed controller model will help self-adaptive software to understand users' need and requirements clearly, so that not only existing parameters could be changed, but also the self-adaptive software can be made to to gain the ability to modify the parameter itself, and more importantly, a good context-aware controlled self-adaptive software could save costs for organisations in addition to improving the effectiveness and efficiency for business in a creative way.