Adaptive Collaborative Air-Ground Edge Caching in Dynamic Vehicular Network Scenarios

Mobile edge caching is a promising technique for supporting the applications of Intelligent Transportation Systems (ITS), which involve huge volumes of image and video contents. Specifically, critical challenges arise in meeting the contents requirements of moving vehicles. In this work, we investigate the collaborative contents delivery in air-ground integrated vehicular networks, where unmanned aerial vehicles (UAVs) and roadside units (RSUs) cache popular contents and further deliver it to nearby vehicles. Efficient contents placement is crucial because of the high mobility of vehicles and UAVs, and the limited caching capacity of edge nodes. To address this, we formulate the heterogeneous cache optimization problem as a generalized multi-dimensional multiple knapsack problem (MMKP), which is NP-hard. In particular, we propose an adaptive collaborative caching optimization (ACCO) algorithm that employs the Kernel Search (KS) framework to find a near-optimal caching scheme, with dynamically adjustable dimensions and solution time for subproblems. Furthermore, we use the k-means clustering algorithm to categorize contents based on service latency, caching cost, and contents size, adding a greedy-based load balancing strategy to achieve cost-effective optimal matching of contents and edge nodes. Comprehensive simulations verify the effectiveness of the proposed cooperative edge caching architecture and the ACCO algorithm in reducing service delay and caching costs for contents delivery.