Computation Offloading for Image Compression in Mobile Edge Computing Using a Deep Belief Network Based on the Markov Approximation Algorithm

The rising popularity of health-related wearables has increased the demand for computational offloading. Latency, battery life, and compute capabilities are all areas that could use significant improvement despite the massive expansion of wearable devices and offloading techniques. There are a number of issues caused by the limited battery life of mobile devices, so it is essential to conserve power by moving responsibilities for running applications to the cloud. This is especially crucial in situations where scheduling is being done in a dynamic mobile cloud computing environment. In this piece, we highlight the fact that most smart wearable devices can pair with smartphones via Bluetooth, and that this connection is significantly more power-efficient than 3G/LTE or Wi-Fi communication. To be more precise, we investigate the origins of this phenomenon. New computing technology, Mobile Edge Computing (MEC), may be the answer to capacity and performance problems in older systems like Mobile Cloud Computing (MCC). Some examples of these problems include excessive latency, a clogged core network, a lack of quality of experience (QoE), and the wasteful consumption of expensive resources like power and data transfer. To address these issues in the MEC setting, we introduced a Deep Belief Network (DBN) equipped with a Markov Approximation Algorithm (MAA). Because of this, a choice could be made that takes into account energy use, time constraints, and system load. AFCFS, MINET, and trade-off judgments for code offloading are all examples of state-of-the-art techniques that can be outperformed by this approach (TRADEOFF). Using entropy encoding as a post-processing step after quantization allows for lossless compression of an image. It allows for a picture to be shown in a way that is both more effective and requires less space for storage or transmission. Our proposed DBN-MAA reduces energy consumption while simultaneously increasing the number of completed jobs, as shown by the simulation experiment results.