Using Spreadsheet-defined Rules for Reasoning in Self-Adaptive Systems

Using rules to capture adaptation knowledge is a common approach for self-adaptive systems. Rule-based reasoning, i.e., using rules to analyze and plan adaptations, has several advantages: (i) it is easy to implement, (ii) it offers fast reasoning, and (iii) it works on resource-spare systems as historical knowledge is not required. Hence, the needed computational power is low and it perfectly suits systems in the pervasive IoT domain. However, the codification of rules poses a challenge to the system design. Existing approaches often require a specific syntax or programming language. Additionally, some approaches force the developer to customize the reasoning mechanism, hence, to reimplement parts of the reasoning. To address these shortcomings, we propose a reusable approach for rule-based reasoning in this paper. Rules can be defined in a spreadsheet without the need to neither learn a syntax nor implement a single line of code. We evaluate the benefits of our approach in two case studies conducted by Master students as well as a quantitative evaluation.