A software energy analysis method using ExecutableUML

The power consumption of an embedded system has been increased. It is important to analyze energy consumption while taking software behaviors into account to reduce energy consumption. Some studies propose model-based energy analysis. Although these method can estimate energy usage within a short time as compared with code-based analysis, the problem is that these technique cannot detect the bottleneck of energy consumption in software behavior in the design stage. This paper proposes an energy consumption analysis method for multi-granularity of ExecutableUML models. The method is useful for detecting the bottleneck of energy consumption in software behavior in the design stage. We also demonstrate that the error of our energy analysis method is, on average, 9.0%.