SMD battery monitoring data

This dataset was generated through an Android application we designed to purposely monitor and record the battery consumption of smart mobile devices (SMDs) while running intensive sorting algorithms. Paper abstract: Computation offloading is an innovative technique that significantly supports the performance of smart mobile devices (SMDs) by leveraging mobile edge computing (MEC) servers’ resources. Considering the current studies under computation offloading and resource allocation, there is a growing emphasis on the development of intelligent and adaptive offloading strategies, nonetheless, these research efforts merely consider the battery state of SMDs and their fast-changing context conditions. To this effect, this paper proposes a deep neural network (DNN) enabled Context-aware Computation Offloading (CaCO) architecture, specifically taking into consideration SMD’s battery consumption when running computation-intensive tasks before suggesting offloads. Furthermore, the formulated task allocation problem is cast as an NP-hard optimization challenge, and to tackle this in accessible time, the paper introduces a solution based on differential evolution (DE) to find the near-optimal solution while employing the DNN model to predict the battery levels of SMDs when in operation. Through extensive experimentation and evaluation in comparison to other baseline algorithms, the proposed framework demonstrates higher performance and potential,showcasing the synergy between DNN for battery level prediction and DE for efficient computation offloading strategy.