Application of linear solvers to UMTS network optimization without and with smart antennas

The paper presents a brief survey of the methods for solving systems of linear equations (linear solvers), which could be useful in application to WCDMA network optimization. The methods are scored with respect to their computational cost when base stations are equipped with traditional antennas as well as with smart antennas. Preconditioning in the conjugate gradient square (CGS) method is discussed. We also analyze the technique, which without any loss of accuracy, may reduce dimensions of the power control problem in a multi-service network but is limited only to traditional antennas the smart antenna case must be solved with conventional solver. Theoretical considerations are confirmed experimentally.