Hybrid Artificial Intelligent Algorithm for Call Admission Control in WCDMA Mobile Network

In wideband code division multiple access (WCDMA) mobile network, total transmission power of Node B depends on diverse factors such as accommodation of new service request, termination of active user equipment (UE) and movement of UE. This makes power prediction a complicated task. In this paper, support vector regression (SVR) has been implemented successfully to forecast next interval power consumption at Node B with different type of antenna system. The predicted output is used by WCDMA mobile network to make decision on new service request admission. Genetic algorithm is then applied to form beams with minimum power to cover all UEs in a macro cell. The proposed algorithm, support vector regression assists genetic algorithm (SVRaGA) was tested in a dynamic WCDMA mobile network simulator. Simulation results have shown SVR can predict next cycle power usage at Node B with excellent accuracy and improve the quality of service (QoS) by minimizing dropped calls in the system.