基于多智能体强化学习的车联网多跳卸载和资源分配方法
By employing the MADDPG algorithm based on multi-agent reinforcement learning and a dynamic agent management mechanism, the problem of dynamic optimization of multi-hop offloading and resource allocation in the Internet of Vehicles (IoV) is solved, achieving low-latency task decision-making and efficient resource allocation, and is suitable for IoV edge computing scenarios.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOUTHEAST UNIV
- Filing Date
- 2026-05-12
- Publication Date
- 2026-07-17
AI Technical Summary
In dynamic vehicle-to-everything (V2X) environments, multi-hop offloading and resource allocation face challenges such as rapid changes in network topology due to high-speed vehicle movement, unstable link connections, difficulties in handling mixed action spaces, and the inability of traditional methods to optimize in real time.
We employ a multi-agent deep deterministic policy gradient (MADDPG) algorithm, combined with Gumbel-Softmax and Sigmoid functions to handle the mixed action space, and design a dynamic agent management mechanism. We adopt a centralized training and distributed execution paradigm to achieve multi-vehicle collaborative decision-making.
It achieves joint optimization of multi-hop offloading and resource allocation in dynamic vehicle networking scenarios, reduces the average service latency of tasks, improves the task completion rate, adapts to network dynamics, and reduces computational complexity.
Smart Images
Figure CN122179841B_ABST