Offloading Revenue Maximization in Multi-UAV-Assisted Mobile Edge Computing for Video Stream

Bin Li, Huimin Shan · 2024

Read Offloading Revenue Maximization in Multi-UAV-Assisted Mobile Edge Computing for Video Stream by Bin Li, Huimin Shan (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-03965.txt.

Introduction

Bin Li and Huimin Shan design a video system in which drones and idle user devices share computing work. Drone-only systems often lack the computing power that live video needs. In their 2024 system, unmanned aerial vehicles pass video tasks to nearby idle devices through device-to-device links. Four choices are tuned together. They are power allocation, video transcoding, computing resource shares, and drone flight paths.

The joint problem is non-convex and hard to solve directly. The team restates it as a Markov decision process. A deep learning method called TD3 then learns good decisions step by step. The goal is the highest profit from tasks passed off across the whole system. Test results show the method beats older algorithms by a clear margin in overall efficiency. The work points to drone systems that earn more from the same fleet.

Catalog record

CREATORBin Li, Huimin Shan
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-03965

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Offloading Revenue Maximization in Multi-UAV-Assisted Mobile Edge Computing for Video Stream Bin Li, Huimin Shan Abstract Traditional video transmission systems assisted by multiple Unmanned Aerial Vehicles (UAVs) are often limited by computing resources, making it challenging to meet the demands for efficient video processing. To solve this challenge, this paper presents a multi-UAV-assisted Device-to-Device (D2D) mobile edge computing system for the maximization of task offloading profits in video stream transmission. In particular, the system enables UAVs to collaborate with idle user devices to process video computing tasks by introducing D2D communications. To maximize the system efficiency, the paper jointly optimizes power allocation, video transcoding strategies, computing resource allocation, and UAV trajectory. The resulting non-convex optimization problem is formulated as a Markov decision process and solved relying on the Twin Delayed Deep Deterministic policy gradient (TD3) algorithm. Numerical results indicate that the proposed TD3 algorithm performs a significant advantage over other traditional algorithms in enhancing the overall system efficiency. I Introduction With the demand for real-time video processing growing significantly, User Devices (UDs) have often failed to meet the real-time requirements when processing advanced video tasks. Mobile Edge Computing (MEC) as an effective solution has been proposed by sinking computing resources to the edge network, through which UDs can offload video tasks to the edge nodes that satisfies local requests without interacting with the core network [ 1 ] . As a result, network congestion is effectively alleviated, and transmission latency can be reduced. In addition, more personalized low-latency video transcoding services are provided to UDs by MEC through real-time understanding of their channel state information [ 2 ] . When it comes to edge nodes, considering the high cost of construction and maintenance, as well as the limited flexibility of small BS servers, Unmanned Aerial Vehicles (UAVs) equipped with MEC servers have been deployed for their incomparable benefits, valued for their high flexibility, high-quality Line-of-Sight (LoS) channel characteristics, low cost, and more [ 3 ] . In harsh environments, computing resources can be deployed in UAVs, where UAVs act as edge computing nodes flying over ground UDs to provide timely computing services, thereby enhance computing capabilities and reduce communication latency. UAVs can be rapidly repositioned to adapt to dynamic environmental changes, making them particularly suitable for emergency situations and distant area service needs [ 4 ] . By introducing UAVs into the MEC network, it is possible to achieve comprehensive optimization of power allocation, video transcoding, computing resource allocation, and UAV trajectory, thereby maximizing system utility [ 5 ] . UAV-assisted Device-to-Device (D2D) communication can assist in resource allocation and coverage extension, improving overall system performance, while ensuring quality of service and communication continuity in dynamic environments or emergency situations [ 6 ] . Although edge computing has significant advantages for real-time tasks processing, it still faces many technical challenges. It is often costly for edge devices to perform tasks, especially since computing intensive tasks increase the energy consumption of the devices, and edge devices may be unwilling to process these computing tasks without rewards. Therefore, establishing an effective incentive mechanism to compensate for energy costs and to attract the active participation of devices is crucial. In addition, the issue of resource allocation is equally critical. Due to the limitation of bandwidth and computing resources, the processing of real-time tasks needs to be flexibly allocated between edge servers and devices. Specifically, tasks may need to be offloaded through dynamically changing links in terms of bandwidth and resources, which poses higher demands on the system’s resource management. Therefore, in edge-based real-time task processing systems, the reasonable incentive mechanisms and resource allocation strategies are key to ensuring efficient system operation [ 7 ] . Existing research has mainly focused on video streaming processing in UAV-assisted MEC networks or promoted real-time video processing by devices through incentive mechanisms. However, it has not fully considered the cost issues of UAVs and other devices when collaboratively processing videos. Additionally, a naturally arising question is whether other UDs and UAVs are willing to undertake computing tasks without compensation. To fill this gap, we propose a UAV-cooperative D2D edge computing task offloading scheme for video streaming, where multiple UAVs collaboratively serve multiple UDs. This scheme leverages idle UDs and multiple UAVs to process video offloading tasks simultaneously and optimizes the video transcoding strategy of UAVs, aiming to maximize the weighted reward. The main technical contributions of this paper are as follows: 1. We present a multi-UAV-assisted MEC network for video streaming. To attract idle UDs and UAVs to participate in video processing, an incentive mechanism is introduced. Videos are transcoded by UAVs based on a dynamic transcoding policy to reduce video processing overhead. By jointly optimizing UAV trajectory, computing resource allocation, and task offloading ratios, a system revenue maximization model is developed. 2. To address the non-convex optimization problem with various practical constraints, it is transformed into a Markov Decision Process (MDP) model. The Twin Delayed Deep Deterministic policy gradient (TD3) algorithm is proposed to find the current optimal task offloading strategy, which optimizes the weighted total utility of the system. 3. A detailed evaluation of the TD3 algorithm is conducted and we verify its significant superiority over existing Deep Reinforcement Learning (DRL) algorithms in maximizing the total utility through numerical experiments. The rest of this paper is organized as follows. Section II provides related work. In Section III, the system model is presented and the optimization problem is formulated. Then, we present the MDP and TD3-based framework, as well as analyze its complexity in Section IV. Simulation results are presented in Section V. Finally, Section VI concludes the paper. II Related Work Currently, serveral studies focus on the task offloading problem in UAV-assisted MEC [ 8 ] , which has gained widespread attention due to their high mobility and easy deployment [ 9 ] . For example, the authors of [ 10 ] investigated the use of UAVs as mobile computing and caching servers for reducing task completion latency. In [ 11 ] , an ultra-reliable low-latency computing offloading system supporting UAVs was proposed to further reduce the latency. The authors of [ 12 ] proposed self-learning strategies for task offloading in UAV networks, providing new solutions for related applications. An optimization method was proposed in [ 13 ] , which significantly increases task offloading and completion rates in multi-UAV MEC systems by jointly optimizing three-dimensional deployment, elevation angle, computing resource allocation, and task offloading. The authors of [ 14 ] optimized the energy consumption and computing latency of the UAV by partially offloading tasks to the UAV in combination with D2D communication models. However, existing work primarily focused on minimizing energy consumption or latency in UAV-assisted MEC systems, which cannot be directly applied to video streaming scenarios. With the continuous advancement of artificial intelligence, video processing is widely used in a variety of services such as motion capture and identification of personnel in disaster areas [ 15 ] . The authors of [ 16 ] studied the multi-user edge-assisted video offloading problem and analyzed it as a game theory problem to be solved by DRL. Due to the limited energy and computing resources of ground UDs [ 17 ] , it is a challenge to

Read the complete text