A DAG-Based Offloading Strategy with Dynamic Parallel Factor Adjustment for Edge Computing in IoV

Area of Science:

• Intelligent transportation systems
• Edge computing
• Network performance optimization

Background:

• Internet of Vehicles (IoV) generates massive data, posing challenges for computing resource allocation.
• Dynamic and real-time IoV task characteristics necessitate adaptive offloading strategies beyond static methods.

Purpose of the Study:

• To propose an adaptive task offloading algorithm for IoV systems.
• To optimize resource utilization and minimize task completion time in dynamic IoV environments.

Main Methods:

• Developed a Dynamic adjustment of Parallel Factor (DPF) algorithm for task offloading.
• Utilized edge computing to dynamically adjust the parallel factor of directed acyclic graphs (DAGs).
• Implemented real-time monitoring of network conditions and vehicle states for dynamic task scheduling.

Main Results:

• The DPF algorithm significantly reduces task delay and improves task success rates compared to static strategies.
• DPF enhances overall system efficiency and resource utilization.
• Experiments show superior performance, especially in high-load scenarios, with markedly better task completion times.

Conclusions:

• The proposed DPF algorithm effectively addresses the limitations of static offloading in IoV.
• Dynamic adjustment of task parallelism is crucial for optimizing IoV systems.
• DPF offers a promising solution for efficient and reliable task offloading in intelligent transportation.