Improved particle swarm optimization and applications to Hidden Markov Model and Ackley function

Particle Swarm Optimization (PSO) is an algorithm based on social intelligence, utilized in many fields of optimization. In applications like speech recognition, due to existence of high dimensional matrices, the speed of standard PSO is very low. In addition, PSO may be trapped in a local optimum. In this paper, we introduce a novel algorithm that is faster and generates superior results than the standard PSO. Also, the probability of being trapped in a local optimum is decreased. To illustrate advantages of the proposed algorithm, we use it to train a Hidden Markov Model (HMM) and find the minimum of the Ackley function.

[1]  Biing-Hwang Juang,et al.  Fundamentals of speech recognition , 1993, Prentice Hall signal processing series.

[2]  Mark A Gregory,et al.  A novel approach for MFCC feature extraction , 2010, 2010 4th International Conference on Signal Processing and Communication Systems.

[3]  Riccardo Poli,et al.  Particle swarm optimization , 1995, Swarm Intelligence.

[4]  Zhen Ji,et al.  A Particle Swarm Optimization for Hidden Markov Model Training , 2006, 2006 8th international Conference on Signal Processing.

[5]  David B. Fogel,et al.  Evolutionary algorithms in theory and practice , 1997, Complex.

[6]  Jun Liu,et al.  A Novel Hybrid PSO-BP Algorithm for Neural Network Training , 2009, 2009 International Joint Conference on Computational Sciences and Optimization.

[7]  Oliver Watts,et al.  Synthesis of Child Speech With HMM Adaptation and Voice Conversion , 2010, IEEE Transactions on Audio, Speech, and Language Processing.

[8]  Yuewen Jiang,et al.  A Modified PSO Algorithm with Line Search , 2010, 2010 International Conference on Computational Intelligence and Software Engineering.

[9]  D. Ackley.  A connectionist machine for genetic hillclimbing , 1987 .