Vehicular edge computing networks (VECNs) can provide a promising solution to support efficient task execution of vehicles. Consider the channel and access time variations caused by the high mobility of vehicles in a vehicular environment when designing task offloading strategies in VECNs. In this paper, we perform multi-path offloading for a task vehicle with serial tasks based on both dynamic communication distances of vehicle-to-infrastructure (V2I) links, that of vehicle-to-vehicle (V2V) links, and slowly varying large-scale fading information of wireless channels. Considering the task vehicle's low delay requirements, our goal is to minimize the maximum task completion time of the task vehicle. A multi-path dynamic offloading scheme (MPDOS), composed of three parts, is proposed to achieve maximum delay minimization. The maximum processing capability of links between a task vehicle and roadside units (RSUs) is first taken as the objective to find the required communication links, which can decrease the total processing time by increasing transmission rate and execution capacity. Then, a task allocation scheme based on a multi-knapsack algorithm matches tasks and RSUs. Finally, a balancing scheme is leveraged to provide load-balancing computing performance across all computation devices. Numerical results show that our proposed scheme outperforms 30.7% of the RA algorithm, and the task completion rate can reach 99.55%.

中文翻译：

---

车载边缘计算网络多路径串行任务卸载策略及动态调度优化


车辆边缘计算网络（VECN）可以提供一种有前途的解决方案来支持车辆高效的任务执行。在设计 VECN 中的任务卸载策略时，请考虑车辆环境中车辆的高移动性引起的信道和访问时间变化。在本文中，我们基于车辆到基础设施（V2I）链路的动态通信距离、车辆到车辆（V2V）链路的动态通信距离以及缓慢变化的大距离，为具有串行任务的任务车辆执行多路径卸载。 -无线信道的尺度衰落信息。考虑到任务车辆的低延迟要求，我们的目标是最小化任务车辆的最大任务完成时间。提出了一种由三部分组成的多路径动态卸载方案（MPDOS），以实现最大延迟最小化。首先以任务车辆和路边单元（RSU）之间链路的最大处理能力为目标来寻找所需的通信链路，这可以通过提高传输速率和执行能力来减少总处理时间。然后，基于多背包算法的任务分配方案匹配任务和RSU。最后，利用平衡方案来提供跨所有计算设备的负载平衡计算性能。数值结果表明，我们提出的方案优于RA算法30.7%，任务完成率可达99.55%。