Hybrid Cooperative Cache Based on Temporal Convolutional Networks in Vehicular Edge Network.

Area of Science:

• Intelligent Transportation Systems
• Wireless Networking
• Network Caching

Background:

• Increasing demand for intelligent vehicle services drives significant wireless network traffic.
• Current edge caching strategies focus on content popularity, leading to redundancy and inefficiency.
• Edge caching offers location advantages for efficient data transmission.

Purpose of the Study:

• To propose a novel hybrid content value collaborative caching strategy (THCS) for edge networks.
• To address cache redundancy and low efficiency in current mainstream caching solutions.
• To optimize cache content and reduce content delivery latency in intelligent vehicle environments.

Main Methods:

• Utilizing Temporal Convolutional Network (TCN) for accurate content popularity prediction.
• Developing a Hybrid Content Value (HCV) metric considering multiple factors for cached content.
• Employing a dynamic programming algorithm for maximizing overall HCV and optimizing cache decisions.

Main Results:

• The proposed THCS strategy demonstrated a 12.3% improvement in cache hit rate compared to benchmark schemes.
• THCS achieved a 16.7% reduction in content transmission delay.
• Simulation experiments validated the effectiveness of the collaborative caching approach.

Conclusions:

• The THCS strategy effectively optimizes edge caching by enabling collaboration between edge nodes.
• Considering hybrid content value alongside popularity enhances caching efficiency and reduces latency.
• This approach provides a viable solution for managing network traffic in the evolving landscape of intelligent vehicles.