Abstract:
Energy consumption analysis and resource allocation (RA) for mobile devices need the
efficient distribution of computational resources and the detailed analysis of power usage
patterns among devices. Using techniques such as predictive modelling, monitoring, energy
consumption patterns, and data collection are examined, enabling informed decisions based on
the RA namely network bandwidth, CPU, and memory. By enhancing RA techniques based on
device workload and real-time energy demands, this method focuses on enhancing energy
effectiveness, extending battery lifetime, and improving overall system performance in mobile
computing environment. This study introduces an innovative approach to monitoring the
energy consumption of mobile devices interconnected to the Raspberry Pi via the web
application interface. Particularly, the focus is on Android mobiles that are wirelessly
connected to the Raspberry Pi through the WiFi network connection. This allows real-time
monitoring of key energy metrics, such as overall energy consumption, CPU usage, and battery
levels, which facilitates informed decision-making based on the RA. Moreover, the Raspberry
Pi applies an XGBoost classifier to efficiently define allocate resources and the idle state of
connected devices based on their usage patterns. The integrated system optimizes energy
efficacy and improves resource utilization, thus contributing to the performance and
sustainability of mobile devices. The system can intelligently allocate resources and predict
device usage based on real-time energy demands through data collection and analysis,
combined with machine learning techniques like XGBoost. The architecture intends to improve
energy efficacy, extend battery lifetime, and improve overall system performance by enhancing
RA, thus contributing to resilient and sustainable mobile computing environments.
