Self-Adaptive Learning of Task Offloading in Mobile Edge Computing Systems

Abstract

Mobile edge computing (MEC) focuses on transferring computing resources close to the user's device, and it provides high-performance and low-delay services for mobile devices. It is an effective method to deal with computationally intensive and delay-sensitive tasks. Given the large number of underutilized computing resources for mobile devices in urban areas, leveraging these underutilized resources offers tremendous opportunities and value. Considering the spatiotemporal dynamics of user devices, the uncertainty of rich computing resources and the state of network channels in the MEC system, computing resource allocation in mobile devices with idle computing resources will affect the response time of task requesting. To solve these problems, this paper considers the case in which a mobile device can learn from a neighboring IoT device when offloading a computing request. On this basis, a novel self-adaptive learning of task offloading algorithm (SAda) is designed to minimize the average offloading delay in the MEC system. SAda adopts a distributed working mode and has a perception function to adapt to the dynamic environment in reality; it does not require frequent access to equipment information. Extensive simulations demonstrate that SAda achieves preferable latency performance and low learning error compared to the existing upper bound algorithms.

Dynamic Offloading Method for Mobile Edge Computing of Internet of Vehicles Based on Multi-Vehicle Users and Multi-MEC Servers

“…In order to reduce the energy consumption of the MEC system to perform tasks, in [22], a heterogeneous two-layer computing offloading framework has been proposed, which can formulate joint offloading and multi-user association problems for multi-user MEC systems and effectively reduce the energy consumption of system processing tasks. Reference [23] proposes a new adaptive offloading algorithm, considering that a large number of resources of mobile devices are underutilized, as well as the spatiotemporal dynamics of devices, the uncertainty of service request volume, and the changes in the communication environment within the MEC system. Reference [24] builds a multi-user MEC system model under channel interference for continuous task execution and data partition-oriented applications.…”

A survey on intelligent computation offloading and pricing strategy in UAV-Enabled MEC network: Challenges and research directions

“…A UAV-MEC-based offloading process consists of three relatively independent processes, i.e., profiling, partitioning, and decision-making [41]:…”

“…Application profiling is the practice of gathering information on programs, such as energy consumption, data size, memory usage, and execution time. MD profiling, on the other hand, collects data on MD status and may not be sent to edge computing, such as battery level, CPU utilization, wireless connection [41], and sojourn time [42]. The sojourn time is the duration of the stay of an MD in UAV-MEC coverage.…”

“…However, unstable spectrum reuse performance can push MDs away, particularly when MDs are used as secondary or cognitive users of cellular networks [155]. The low expectations for D2D services and frequent mode switching can obstruct successful D2D offloading, cause severe network congestion, and result in poor QoE, particularly in densely populated areas [41]. Many authors focus on how to improve the use of D2D by encouraging the MD with idle resources to share its resource at a good price [156] [157].…”