Abstract
Intelligent vehicles have provided a variety of services; there is still a great challenge to execute some computing-intensive applications. Edge computing can provide plenty of computing resources for intelligent vehicles, because it offloads complex services from the base station (BS) to the edge computing nodes. Before the selection of the computing node for services, it is necessary to clarify the resource requirement of vehicles, the user mobility, and the situation of the mobile core network; they will affect the users’ quality of experience (QoE). To maximize the QoE, we use multiagent reinforcement learning to build an intelligent offloading system; we divide this goal into two suboptimization problems; they include global node scheduling and independent exploration of agents. We apply the improved Kuhn–Munkres (KM) algorithm to node scheduling and make full use of existing edge computing nodes; meanwhile, we guide intelligent vehicles to the potential areas of idle computing nodes; it can encourage their autonomous exploration. Finally, we make some performance evaluations to illustrate the effectiveness of our constructed system on the simulated dataset.
Highlights
With the rapid development of intelligent vehicles, the vehicular network based on artificial intelligence has attracted extensive attention; its wide application has encouraged researchers all over the world to develop more applications, but there is still a problem to compute-intensive services on vehicles; as a promising solution, mobile edge computing (MEC) lets users upload services to edge computing servers (e. g., offloading), which can reduce the computing load of the terminal, just like roadside unit (RSU), building cloud, and other entities with computing [1]
We focus on the tradeoff between the quality of experience (QoE) of users and the profit of servers [7]; we need to schedule the corresponding edge computing nodes for intelligent vehicles; this process is similar to the order dispatch in modern taxi networks [8,9,10]; they provide information about passenger demand and taxi movement for finding the most appropriate pairs; some car-hailing services provide significant improvements over traditional taxi systems in terms of reducing taxi cruising time and waiting time [10, 11]; the online car-hailing is a vivid scene [8, 12,13,14] which can be migrated to the edge computing
About (2), we show that both the results of traditional and Multiagent recurrent deep deterministic policy algorithm (MARDDPG) under the same RSU and different numbers of vehicles, it is used to compare different algorithms and highlight the performance under different sparsity levels of the agent. e use of different metrics is to show the performance of each algorithm, it is difficult to finish this matching problem in limited time, it is obvious that each algorithm has its limitations and merits
Summary
With the rapid development of intelligent vehicles, the vehicular network based on artificial intelligence has attracted extensive attention; its wide application has encouraged researchers all over the world to develop more applications, but there is still a problem to compute-intensive services on vehicles; as a promising solution, mobile edge computing (MEC) lets users upload services to edge computing servers (e. g., offloading), which can reduce the computing load of the terminal, just like roadside unit (RSU), building cloud, and other entities with computing [1]. G., offloading), which can reduce the computing load of the terminal, just like roadside unit (RSU), building cloud, and other entities with computing [1]. The performance of traditional methods [2, 3] will decline sharply in the vehicular network; it is urgent to develop an effective MEC intelligent offloading solution. Machine learning is mainly used in intensive computation tasks, such as navigation and automatic driving, except in MEC; it is very difficult to build a suitable model because many vehicles are participating in the offloading system. DRL has been applied to the MEC offloading system [4], mainly for task scheduling and resource allocation and for studying DRL-based networking and caching [5, 6]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
