LAUSR

Vertical handover decision management is an essential to keep the seamless ubiquitous heterogeneous wireless networks since each access network has different operations in the next generation. In this paper, WCDMA, LTE and WLAN are cooperated in the architecture of mobile IP regional registration (MIP-RR). We propose the Learning Vector Quantization Neural Networks (LVQNNs) approach in order to maintain uninterrupted communication that depends on received signal strength indicator, data rate requirement, monetary cost of service and mobile terminal device speed metrics are considered as multi-criteria to initial handover. Furthermore, the multi-criteria are dynamic to influence for real-time and non-real time services in different networks and the users select the optimal target network which is the highest handover factor score in order to balance against the network condition and user preference. To ensure the Always Best Connected (ABC) demands, the simulation results illustrate that our proposed algorithm provided outperform the performance in term of unnecessary handover, the call dropping probability, data packet delay and network utilization compared with conventional method as fuzzy logic and neural network based machine learning.