Learning-Based Scheme for Efficient Content Caching in Vehicular Networks

Abstract

AbstractIn vehicular ad hoc networks (VANETs), caching is a very promising technique to reduce the transmission burden and to improve the users’ Quality of Experience (QoE) in terms of latency. Increasing cache hit ratio is very important for delay sensitive applications. In this paper, average cache hit ratio maximization problem is proposed and formulated while taking into account the time-varying topology of network, erratic vehicular (user) mobility, varying requests and preferences of multiple users and the limited cache capacity of the Road Side Units (RSUs). A learning automata-based cache update policy has been designed in order to determine appropriate content to be cached in RSUs. The performance of the learning scheme-based caching policy has been evaluated using simulations and analysed in comparison with three other caching policies. Simulation results indicate that the learning-based caching policy can significantly improve the average cache hit ratio, minimize latency, and thus, enhance the Quality of Experience.