Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Computation Offloading for Multi-server Multi-access Edge Vehicular Networks: A DDQN-based Method

Siyu Wang, Bo Yang, Zhiwen Yu, Xuelin Cao, Yan Zhang, Chau Yuen

TL;DR

This work addresses computation offloading in a multi-server MEC for vehicular networks where terminals are mobile and tasks have priority. It tackles the problem in two stages: an offloading decision stage that weighs delay and energy benefits, and a DDQN-based offloading-request scheduling stage that optimally allocates MEC resources under mobility-induced dynamics. The proposed R-DDQN algorithm Contextualizes server load and mobility to schedule tasks by priority, surpassing traditional mathematical methods and standard DQN in simulations. The approach demonstrates improved timely processing of important tasks and robust performance as the number of devices and mobility rates grow, offering practical gains for low-latency, mobility-aware edge computing in connected-vehicle environments.

Abstract

In this paper, we investigate a multi-user offloading problem in the overlapping domain of a multi-server mobile edge computing system. We divide the original problem into two stages: the offloading decision making stage and the request scheduling stage. To prevent the terminal from going out of service area during offloading, we consider the mobility parameter of the terminal according to the human behaviour model when making the offloading decision, and then introduce a server evaluation mechanism based on both the mobility parameter and the server load to select the optimal offloading server. In order to fully utilise the server resources, we design a double deep Q-network (DDQN)-based reward evaluation algorithm that considers the priority of tasks when scheduling offload requests. Finally, numerical simulations are conducted to verify that our proposed method outperforms traditional mathematical computation methods as well as the DQN algorithm.

Computation Offloading for Multi-server Multi-access Edge Vehicular Networks: A DDQN-based Method

TL;DR

This work addresses computation offloading in a multi-server MEC for vehicular networks where terminals are mobile and tasks have priority. It tackles the problem in two stages: an offloading decision stage that weighs delay and energy benefits, and a DDQN-based offloading-request scheduling stage that optimally allocates MEC resources under mobility-induced dynamics. The proposed R-DDQN algorithm Contextualizes server load and mobility to schedule tasks by priority, surpassing traditional mathematical methods and standard DQN in simulations. The approach demonstrates improved timely processing of important tasks and robust performance as the number of devices and mobility rates grow, offering practical gains for low-latency, mobility-aware edge computing in connected-vehicle environments.

Abstract

In this paper, we investigate a multi-user offloading problem in the overlapping domain of a multi-server mobile edge computing system. We divide the original problem into two stages: the offloading decision making stage and the request scheduling stage. To prevent the terminal from going out of service area during offloading, we consider the mobility parameter of the terminal according to the human behaviour model when making the offloading decision, and then introduce a server evaluation mechanism based on both the mobility parameter and the server load to select the optimal offloading server. In order to fully utilise the server resources, we design a double deep Q-network (DDQN)-based reward evaluation algorithm that considers the priority of tasks when scheduling offload requests. Finally, numerical simulations are conducted to verify that our proposed method outperforms traditional mathematical computation methods as well as the DQN algorithm.
Paper Structure (13 sections, 15 equations, 6 figures)

This paper contains 13 sections, 15 equations, 6 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. Overview
  4. Terminal Device Model
  5. MEC Server Model
  6. Problem Formulation and Analysis
  7. Problem formulation
  8. Problem Analysis
  9. Offloading decision making
  10. Offloading request scheduling
  11. Tasks uploading
  12. Experimental Results
  13. Conclusion

Figures (6)

  • Figure 1: The considered multi-server multi-access edge vehicular network scenario, where the terminal device (e.g., the blue vehicle) within the overlapped area needs to choose the optimal offloading strategy.
  • Figure 2: Terminal device model.
  • Figure 3: MEC server model.
  • Figure 4: The proposed R-DDQN algorithm flow.
  • Figure 5: The average loss in different learning rates is shown in (a), the average loss in different batch sizes is shown in (b).
  • ...and 1 more figures