Hierarchical Matching With Peer Effect for Low-Latency and High-Reliable Caching in Social IoT

The Internet of Things (IoT) is expected to bring great benefits to users, operators, and manufactures in different application scenarios. Given that smart objects which can exploit their own social networks and share contents via device-to-device (D2D) communications, the combined social IoT promises to collect information as well as to provide services more efficiently. This application scenario requires lower latency and higher reliability, and then we adopt D2D-based caching to reduce the downloading latency while guaranteeing the reliable delivery. To achieve this joint optimization objective, we conceive the interdependence with the framework of hierarchical bipartite graph. In this way, this combinatorial problem can be decoupled into a content sharing problem and a resource allocation problem. To solve the first one, we propose a content sharing-oriented matching algorithm with projecting social characters onto physical links. To solve the second one, we formulate this resource allocation problem as equivalent to a many-to-one matching game with peer effect. We then design a novel distributed algorithm with rotation-swap, which can converge to a stable state with limited number of iterations. Formulating the convergence procedure as another NP-hard problem, we further design a coloring-based heuristic algorithm to find a near-optimal solution. We conduct extensive simulations to demonstrate that our proposed schemes can achieve a better tradeoff between performance and complexity than other benchmarks.