Abstract

Abstract-In the era of Internet of Things, software systems need to interact with many physical entities and cope with new requirements at runtime. Self-adaptive systems aim to tackle those challenges, often representing their context with a runtime model enabling better reasoning capabilities. However, those models quickly grow in size and need to be updated frequently with small changes due to a high number of physical entities changing constantly. This situation threatens the efficacy of analyses on such models, as they lack an efficient management of those changes leading to unnecessary computation overhead. We propose applying scalable, incremental change management of runtime models in the presence of a complex model to text transformation. In this paper, we present and evaluate an example of code generation of integer linear programs. In our case study using synthesized models, we saved 35 -83% processing time compared to a non-incremental approach. Using our approach, future self-adaptive systems can handle and analyze large-scale runtime models, even if they change frequently.