A Weighted Voting Mechanism for Action Selection Problem in Self-Adaptive Software

Self-adaptive software is a closed-loop system which aims at adjusting itself in response to changes at runtime. Such a system is required to monitor domain events, detect significant changes, decide how to react, and act in order to execute the decisions. This paper focuses on the deciding process particularly for application-level adaptation actions. For this purpose, a weighted voting mechanism has been proposed which makes decisions based on a Goal-Action- Attribute Model (GAAM). The decision-making algorithm traverses GAAM, determines activated goals and feasible actions for voting, and ultimately selects an action as the social choice. The proposed mechanism and GAAM are evaluated within a simulated model of a news web site.