Title :
Application of reinforcement learning to admission control in CDMA network
Author :
Makarevitch, Boris
Author_Institution :
Commun. Lab., Helsinki Univ. of Technol., Espoo, Finland
Abstract :
The paper describes an admission control algorithm for the CDMA networks which is able to adapt to the operating environment. The algorithm is based on the principle of reinforcement learning and it achieves near-optimal performance for various radio propagation conditions and network operator´s objectives. The performance evaluation results for different state space alternatives and algorithm parameters are presented and compared with the conventional admission control based on the power thresholds
Keywords :
Markov processes; cellular radio; code division multiple access; learning (artificial intelligence); multiuser channels; radio networks; telecommunication congestion control; CDMA network; Markov decision process; admission control algorithm; algorithm parameters; cellular radio; near-optimal performance; network operator objectives; performance evaluation results; power thresholds; radio propagation conditions; reinforcement learning; state space alternatives; Admission control; Base stations; Communication system control; Degradation; Downlink; Intelligent networks; Interference; Learning; Multiaccess communication; Paper technology;
Conference_Titel :
Personal, Indoor and Mobile Radio Communications, 2000. PIMRC 2000. The 11th IEEE International Symposium on
Conference_Location :
London
Print_ISBN :
0-7803-6463-5
DOI :
10.1109/PIMRC.2000.881639
