Methodological Considerations for Self-adaptive Systems: An Essay

TL;DR

The paper addresses methodological questions for uncertainty and risk-aware adaptation in self-adaptive systems and proposes architecture-based approaches to manage runtime adaptation. It adopts a design-science methodology with nested problem decomposition to connect practical problems and knowledge questions. Three empirical methods—systematic literature reviews, surveys, and case studies—are used to map questions to methods and validate the proposed viewpoint. The contributions include a reusable viewpoint for balancing adaptation benefits, costs, and risks across the lifecycle, demonstrated via a DeltaIoT IoT case.

Abstract

In this essay, we provide an overview of methodological considerations necessary to lay out the foundation for our PhD research on uncertainty and risk-aware adaptation.

Figures (3)

• Figure 1: The seven waves of research on engineering self-adaptive systems weyns2020book.
• Figure 2: A simplified overview of the design science framework by Wieringa.
• Figure 3: The decomposition of the design problem tackled in this dissertation.