Call admission control using support vector machine for policy-based QoS control

摘要

A call admission control algorithm using Support Vector Machine (SVM) (SVM-CAC) is analyzed. SVM-CAC use the service vector and network vector to predict admission state. QoS metric function compares with some thresholds to determine the admission state. The thresholds value can reveal biases of services. SVM-CAC combines policy and SVM's advantages when make admission decision, so it can take into account the business requirement, external network QoS resource and reduce algorithm complexity. The simulation results show that this scheme accelerates calculation speed, have lower call delay, achieve superior performance in terms of the call blocking probability and the call dropping probability than other machine learning admission control.