International Journal of Electrical and Computer Engineering (IJECE) Vol. 15, No. 2, April 2025, pp. 1924~1932 ISSN: 2088-8708, DOI: 10.11591/ijece.v15i2.pp1924-1932  1924 Journal homepage: http://ijece.iaescore.com Next-generation offloading using hybrid deep learning network for adaptive mobile edge computing P. Anusha, V. Mary Amala Bai Department of Information Technology, Noorul Islam Centre for Higher Education, Kanyakumari, India Article Info ABSTRACT Article history: Received May 10, 2024 Revised Sep 19, 2024 Accepted Oct 1, 2024 Deploying mobile application tasks that require a lot of computing and are time-sensitive to distant cloud-based data centers has become a popular method of working around the limitations of mobile devices (MDs). Deep reinforcement learning (DRL) techniques for offloading in mobile edge computing (MEC) environments struggle to adapt to new situations due to low sample efficiency for each new context. To address these issues, a novel computational offloading in mobile edge computing (COOL-MEC) algorithm has been proposed that combines the benefits of attention modules and bi-directional long short-term memory. This algorithm improves server resource utilization by lowering the cost of assimilating processing latency, processing energy consumption, and task throughput of latency-sensitive tasks. The experiment's findings show that, when used as intended, the recommended COOL-MEC algorithm minimizes energy consumption. When compared to the current deep convolutional attention reinforcement learning with adaptive reward policy (DCARL-ARP) and DRL techniques, the energy consumption of the proposed COOL-MEC is decreased by 0.06% and 0.08%, respectively. The average time per channel utilized for the execution of the proposed COOL-MEC also decreased by 0.051% and 0.054% when compared with existing DCARL-ARP and DRL methods respectively. Keywords: Average energy consumption Computation offloading Deep learning Deep reinforcement learning Long short-term memory Mobile devices Mobile edge computing This is an open access article under the CC BY-SA license. Corresponding Author: P. Anusha Department of Information Technology, Noorul Islam Centre for Higher Education Kumaracoil, Thucklay, Kanyakumari, Tamil Nadu, 629 180, India Email: mail2meanu82@gmail.com 1. INTRODUCTION The internet of things (IoT) smart mobile device (SMD) is a powerful computing device that enables smart networking [1], [2]. An emerging technology called the IoT enables real objects, including cars and home appliances, to interact and even converse with one another [3]–[5]. Simultaneously, SMDs are frequently used to implement applications like virtual reality and interactive online gaming that demand supercomputing capacity, extremely low latency, and perpetual access rights [6], [7]. On the other hand, given that SMDs are portable, their small size results in higher energy consumption, lower processing power, and smaller storage capacity [8]. Many apps are exceedingly difficult to deploy because of this SMD constraint [9]. To overcome these restrictions, SMD uses a wireless network to connect to a remote cloud and moves computational functions to the cloud [10]. However, most sizable data centers housing cloud computing resources remotely are situated at a considerable distance from the bulk of clients [11]. Because of this, SMD will take longer to unload and consume more energy during the process [12]. Mobile edge computing, or MEC, has been developed as a recent solution to the previously described problems. Edge servers, frequently referred to as compute nodes, are dispersed throughout the network when using MEC [13]. Edge servers are thus situated closer to users than independent cloud servers and are capable