Uncertainty in self-adaptive software systems

Abstract. The ever-growing complexity of software systems coupled with their stringent availability requirements are challenging the manual management of software after its deployment. This has motivated the development of self-adaptive software systems. Self-adaptation endows a software system with the ability to satisfy certain objectives by au-tomatically modifying its behavior at runtime. While many promising approaches for the construction of self-adaptive software systems have been developed, the majority of them ignore the uncertainty underlying the adaptation. This has been one of the key inhibitors to widespread adoption of self-adaption techniques in risk-averse real-world applica-tions. Uncertainty in this setting is a vaguely understood term. In this paper, we characterize the sources of uncertainty in self-adaptive software system, and demonstrate its impact on the system’s ability to satisfy its objectives. We then provide an alternative notion of optimality that ex-plicitly incorporates the uncertainty underlying the knowledge (models) used for decision making. We discuss the state-of-the-art for dealing with uncertainty in this setting, and conclude with a set of challenges, which provide a road map for future research