Multi-agent caching distribution strategy for content freshness guarantee in IoV

Abstract

Vehicles need to dynamically changing content to support latency-sensitive applications in Internet of vehicles (IoV), thereby increasing the load on the macro base station (MBS) and reducing the freshness of content. Utilizing edge caching to cache the latest content in small base station (SBS) can effectively reduce the latency and improve the content freshness. An in-depth analysis was conducted on latency and content's age of information (AoI). A content freshness assurance multi-agent reinforcement learning (MARL) algorithm was proposed, which optimized cache distribution decisions to guarantee high freshness. Simulation results show that the proposed algorithm not only converges faster but also demonstrates better performance in reducing latency and enhancing content freshness.

Keywords

• Internet of vehicles
• edge caching
• age of information
• multi-agent reinforcement learning