Search

With the rapid advancement of Internet of vehicles and autonomous driving technology, there is a growing need for increased computing power in vehicle operations. However, the strict latency requirements of vehicle tasks may pose challenges to communication and computing resources within the vehicle edge computing network. This paper introduces a two-stage joint optimization to address challenges, focusing on minimizing vehicle task latency and optimizing resource allocation. In addition, the task completion rate is considered an important indicator to ensure safety and reliability in practical application scenarios. Next, we propose a global adaptive offloading and resource allocation optimization model named GOAL. The GOAL model dynamically adjusts the weight coefficients of the reward function to optimize the model, integrating the actor-critic algorithm to effectively adapt to uncertain environments. Through experimental comparisons of various weight coefficients for task arrival rates and reward functions, we were able to determine the optimal hyperparameters for our proposed model. The simulation results show that the GOAL model outperforms the benchmark methods by over 30% in reward value. It also performs better in terms of task delay and energy consumption. Additionally, the GOAL model has a higher task completion rate compared to the benchmark methods, and it exhibits strong search capabilities and faster convergence speed. [ABSTRACT FROM AUTHOR]

Cloudlet-based vehicular networks are a promising paradigm to enhance computation services through a distributed computation method, where the vehicle edge computing (VEC) cloudlet are deployed in the vicinity of the vehicle. In order to further improve the computing efficiency and reduce the task processing delay, we present a parallel task scheduling strategy based on the multi-agent deep reinforcement learning (DRL) approach for delay-optimal VEC in vehicular networks, where multiple computation tasks select the target threads in a VEC server to execute the computing tasks. We model the target thread decision of computation tasks as a multi-agent reinforcement learning problem, which is further solved by using a task scheduling algorithm based on multi-agent DRL that is implemented in a distributed manner. The computation tasks, with each selection acting on the target thread acting as an agent, collectively interact with the VEC environment and receive observations with respect to a common reward and learn to reduce the task processing delay by updating the multi-agent deep Q network (MADQN) using the obtained experiences. The experimental results show that the proposed DRL-based scheduling algorithm can achieve significant performance improvement, reducing the task processing delay by 40% and increasing the processing probability of success for computation tasks by more than 30% compared with the traditional task scheduling algorithms. [ABSTRACT FROM AUTHOR]

Mobile Edge Computing (MEC) is a promising technology to extend the diverse services to the edge of Internet of Things (IoT) system. However, the static edge server deployment may cause “service hole” in IoT networks in which the location and service requests of the User Equipments (UEs) may be dynamically changing. In this paper, we firstly explore a vehicle edge computing network architecture in which the vehicles can act as the mobile edge servers to provide computation services for nearby UEs. Then, we propose as vehicle-assisted offloading scheme for UEs while considering the delay of the computation task. Accordingly, an optimization problem is formulated to maximize the long-term utility of the vehicle edge computing network. Considering the stochastic vehicle traffic, dynamic computation requests and time-varying communication conditions, the problem is further formulated as a semi-Markov process and two reinforcement learning methods: $Q$ -learning based method and deep reinforcement learning (DRL) method, are proposed to obtain the optimal policies of computation offloading and resource allocation. Finally, we analyze the effectiveness of the proposed scheme in the vehicular edge computing network by giving numerical results. [ABSTRACT FROM AUTHOR]