论文信息 - Convergence Analysis of a GA-ICA Algorithm

Convergence Analysis of a GA-ICA Algorithm

In this work we consider the extension of Genetic-Independent Component Analysis Algorithms (GA-ICA) with guiding operators and prove their convergence to the optimum. This novel method for Blindly Separating unobservable independent component Sources (BSS) consists of novel guiding genetic operators (GGA) and finds the separation matrix which minimizes a contrast function. The convergence is shown under little restrictive conditions for the guiding operator: its effect must disappear in time like the simulated annealing.

[1] Erkki Oja,et al. Independent component analysis: algorithms and applications , 2000, Neural Networks.

[2] David Aaron,et al. A New Concept , 1974 .

[3] Jun Wang,et al. Nonlinear Blind Source Separation Using Higher Order Statistics and a Genetic Algorithm , 2001 .

[4] Juan Manuel Górriz,et al. Hybridizing Genetic Algorithms with ICA in Higher Dimension , 2004, ICA.

[5] Allan Kardec Barros,et al. Independent Component Analysis and Blind Source Separation , 2007, Signal Processing.

[6] Sergio Cruces,et al. Robust blind source separation algorithms using cumulants , 2002, Neurocomputing.

[7] Pierre Comon,et al. Independent component analysis, A new concept? , 1994, Signal Process..

[8] Juan Manuel Górriz,et al. New Model for Time-Series Forecasting Using RBFs and Exogenous Data , 2003 .

[9] Chrystopher L. Nehaniv,et al. Linear Analysis of Genetic Algorithms , 1998, Theor. Comput. Sci..

[10] Juan Manuel Górriz,et al. A new model for time-series forecasting using radial basis functions and exogenous data , 2004, Neural Computing & Applications.