Using search-based software engineering to handle the changes with uncertainties for self-adaptive systems

The changes confronting contemporary Self-Adaptive Systems (SASs) are characterized by uncertainties in their relationships, priorities, and contexts. To generate adaptation strategies for handling these changes, existing adaptation planning methods, which ignore these uncertainties, must be improved. This thesis explores the possibilities of using Search-Based Software Engineering (SBSE) to establish a search-based planning method capable of handling multiple changes in an uncertain context without defining their priorities. Meanwhile, both the assurance approach to improving the efficiency of adaptation planning and the selection approach to choosing a unique strategy are proposed to solve emerging research questions that arise when such planning method is applied in actual SASs. From this experience, we are able to derive innovative methods for the designers of SASs as a reference, which may observably improve the ability of SASs and promote the widespread use of SBSE in SASs.