Q-learning based vertical handover decision algorithm in LTE-A two-tier macrocell-femtocell systems / Ammar Bathich

Next generation mobile systems require improved capacity and proper quality assurance for services. To achieve these requirements, small cells have been deployed intensively by Long Term Evolution (LTE) networks operators beside conventional base station structure to provide customers with better service and capacity coverage. Accomplishment of seamless handover between macrocell layer (first tier) and femtocell layer (second tier) is one of the key challenges to attain the Quality of Service (QoS) requirements. Handover related information gathering becomes very hard in high dense femtocell networks. Effective handover decision designs are important to minimize unnecessary handovers from occurring and avoiding the ping-pong effect. In this work, we have classified and reviewed current works for the two-tier macrocell-femtocell LTE-A systems, based on the main decision scheme applied. Key features and drawbacks of the most representative approaches have been studied. Modeling issues and the key performance evaluation for HeNB-specific mobility management are discussed and studied as well. The main objective of this work is to propose and implement an efficient handover decision procedure based on users’ profiles using Q-learning technique in a LTE-A macrocell-femtocell networks. New multi-criterion handover decision parameters are proposed in typical/dense femtocells in macrocells environment to estimate the target cell for handover. The proposed handover algorithms are validated using the LTE-Sim simulator under an urban environment. The simulation results are encouraging. A reduction in the average number of handovers, the average number of control signaling measurements and packet loss and delay is observed. At the same time, the system throughput was increased. The proposed algorithms reduced the handover number, packet delay and packet loss by 48%, 89% and 85% respectively, whereas the system throughput increased by 13% compared to reference algorithms.