Hariri

Self-adaptive systems (SAS) automatically mitigate environmental changes and unexpected system issues at run time by adapting towards optimal configurations that enable continual requirements satisfaction. The increasing proliferation of SASs presents engineering challenges that reflect issues experienced by non-adaptive systems, more specifically, ensuring that continuing assurance for software artifacts is provided. In particular, ensuring that requirements traceability links are appropriately managed at run time in SASs can be an error-prone procedure and may require significant effort from a requirements engineer. Natural language processing (NLP) techniques have been used to recover broken or missing traceability links efficiently between requirements and other artifacts, however, performing traceability link recovery can introduce significant overhead for SASs. Specifically, the state-space explosion of possible combinations of environmental states, system parameters, and expressed behaviors can lead to states in which no traceability link exists, thereby necessitating recovery. This paper proposes Adaptive Requirements Traceability (ART), a conceptual framework for handling traceability recovery in terms of SASs. We motivate this framework with an illustrative example in the networking domain.