A novel particle-pair cooperative optimizer(PPCO)is proposed for speaker identification based on vector quantization.In this algorithm

two initial particle-pairs simultaneously explore for the best speaker codebook

and each particle-pair which consists of two particles performs basic operations of particle swarm optimization(velocity updating and position updating)and conventional LBG algorithm in sequence at each iteration.Information is exchanged when particle-pairs are reorganized periodically.And then two elitist particles selected from two initial particle-pairs respectively continue to move toward the global optimum.Experimental results have demonstrated that the performance of this new algorithm is much better than that of LBG

FCM

FRLVQ-FVQ

FEP and PSO consistently with lower speaker identification error rates

shorter computational time and higher convergence rate.The dependence of the final codebook on the selection of the initial codebook is also reduced effectively.