Deep Reinforcement Learning-Based Distributed Collaborative Service Caching in Edge Computing
Wei Yang, Yi Chu, Chao Chen, Shengming Jiao, Xiaolong Xu, Shengjun Xue · 2024
As the Internet of Things (IoT) continues to evolve, smart devices increasingly depend on edge computing for services characterized by low latency and high efficiency. The rapid proliferation of smart devices and the diversification of service requests impose greater demands on the quality of edge service provision. Traditional caching strategies based on service request volumes, which neglect the geographic distribution of devices and the randomness of requests, fail to meet the dual requirements of accuracy and response speed, resulting in inefficient resource allocation and caching. To tackle these challenges, this paper introduces ECPG, a novel distributed collaborative service caching strategy leveraging deep reinforcement learning. By modeling the service caching issue as a Markov Decision Process (MDP) and optimizing it through the Deep Deterministic Policy Gradient (DDPG) algorithm, this strategy facilitates independent yet collaborative caching decisions by devices based on real-time data and environmental conditions. This approach significantly improves service response times and overall system efficiency. Experimental results demonstrate ECPG's superior performance in reducing service latency and increasing cache hit rates, underscoring its practicality and efficacy in contemporary edge computing environments.