CAWO: Congestion-aware WiFi offloading for 5G heterogeneous wireless network

The unprecedented data increase has imposed great challenges to cellular networks. Traffic offloading by heterogeneous radio access technologies (RATs) is an effective approach for enhancing network capacity. Without costly and time-consuming infrastructure investments, WiFi offloading is deemed as the prospective evolution for the heterogeneous integration. However, previously proposed schemes mainly focus on alleviating burden by offloading data to WiFi as much as possible, without systematic considerations of the network congestion. In this paper, based on Markov process, we investigate the congestion offloading problem in user random mobility model, and prove it could be solved in subgradient method with equivalent transformation. Moreover, the congestion-aware WiFi offloading algorithm is proposed for multiple WiFi APs network offloading scenario, aiming to balance the capacity increase with congestion characterized by blocking probability. In addition, the upper bound and lower bound of blocking probability is deduced in optimization. Simulation results show that, by optimizing user offloading probability, the algorithm proposed could achieve maximum throughput with lower blocking probability.