Using Reference Attribute Grammar-Controlled Rewriting for Energy Auto-Tuning

Cyber-physical systems react on events reported by sensors and interact with objects of the real world according to their current state and their view of the world. This view is naturally represented by a model which is continuously analysed and updated at runtime. Model analyses should be ideally concise and efficient, requiring well-founded, comprehensible implementations with efficient reasoning mechanisms. In this paper, we apply reference attribute grammar controlled rewriting to concisely implement the runtime model of an auto-tuning case study for energy optimization. Attribute functions are used to interactively perform analyses. In case of an update, our system incrementally—and, thus, efficiently—recomputes depending analyses. Since reference attribute grammar controlled rewriting builds the required dependency graphs automatically, incremental analysis comes for free.

• Type: Journal
• Year: 2015
• Where: MODELS
• Authors: Christoff Bürger, Johannes Mey, René Schöne, Sven Karol, Daniel Langner
• Keywords: MODELS 2015, Natural Language Processing