Abstract: An amended *LMS* Newton algorithm is proposed which create the nonlinear functional relation of the step factor μ and the error signal e (n). And the next we discuss the factors in the amended algorithm and the performance in the different IP environment. The algorithm is simple and easy to be implemented, Theoretical analysis and computer simulation show that the properties of the algorithm such as convergence speed, and steady state error ate better than those of *SVSLMS *algorithm.

458

Authors: Feng Zhang, Yi Nong Li, Lei Wang

Abstract: In active vibration control, most of the signals to be eliminated are deterministic signals. Simulations are carried out using periodic sinusoidal reference signal based on FxLMS algorithm. Based on a deterministic reference signal input conditions, the paper provides a detailed behavior analysis due to the secondary-path model error and phase delay, the results will give theoretical basis by adjusting the FxLMS algorithm parameters to ensure convergence and to achieve the best control performance in practical applications.

3262

Authors: Na Li, Yuan Xiang Li

Abstract: A new particle swarm optimization algorithm (a diversity guided particles swarm Optimization), which is guided by population diversity, is proposed. In order to overcome the premature convergence of the algorithm, a metric to measure the swarm diversity is designed, the update of velocity and position of particles is controlled by this criteria, and the four sub-processes are introduced in the process of updating in order to increase the swarm diversity, which enhance to the ability of particle swarm optimization algorithm (PSO) to break away from the local optimum. The experimental results exhibit that the new algorithm not only has great advantage of global search capability, but also can avoid the premature convergence problem effectively.

1429

Authors: Heng Yang, Jing Wang, Jing Guan, Wei Lu

Abstract: The traditional constant modulus algorithm (CMA) has the disadvantage of slow convergence in blind equalization algorithm. This paper studied one improved algorithm based on momentum factor constant modulus algorithm(MCMA) to solve this problem, momentum factor was added to the weight vector iteration formula of CMA to improve the convergence speed. theoretical analysis and simulation showed that: in the case of the same equalization effect, the MCMA converges faster than the traditional constant modulus algorithm, and also different momentum factors have different convergence effects. The larger the momentum factor , the better convergence effect in the defined domain of the momentum factor.

1058

Authors: Ying Ying Sun, Wei Pei, Li Kong

Abstract: In order to compute and analyze power flow of a distribution network with distributed generations(DGs) and eliminate various problems caused by placing of the DGs, several improved Newton methods were proposed successively, and initial value choice is one of the non negligible issues. The traditional initial value choice method also cannot ensure convergence under some circum stances, and the convergence problem is aggravated by the placing of the DGs. For the convergence problem in the computation for the distribution network with vast DGs accessed by the Newton method, this article proposes an approximation algorithm based on series expansion for computing voltage approximation as the computed initial value of power flow so as to improve algorithm convergence. Furthermore, effects of the number of the DGs, node types, their locations, etc. on error between the voltage approximation and the exact voltage value are analyzed through examples. Example verification shows that the error between the voltage approximation computed by the method described in this article and the obtained final value through calculation is small, and replacing the traditional initial value with the voltage approximation can effectively improve convergence of the power flow computation and reduce iterations.

2193