The Novel Approach of Adaptive Twin Probability for Genetic Algorithm

The performance of GA is measured and analyzed in terms of its performance parameters against variations in its genetic operators and associated parameters. Since last four decades huge numbers of researchers have been working on the performance of GA and its enhancement. This earlier research work on analyzing the performance of GA enforces the need to further investigate the exploration and exploitation characteristics and observe its impact on the behavior and overall performance of GA. This paper introduces the novel approach of adaptive twin probability associated with the advanced twin operator that enhances the performance of GA. The design of the advanced twin operator is extrapolated from the twin offspring birth due to single ovulation in natural genetic systems as mentioned in the earlier works. The twin probability of this operator is adaptively varied based on the fitness of best individual thereby relieving the GA user from statically defining its value. This novel approach of adaptive twin probability is experimented and tested on the standard benchmark optimization test functions. The experimental results show the increased accuracy in terms of the best individual and reduced convergence time.

SVD-EBP Algorithm for Iris Pattern Recognition

This paper proposes a neural network approach based on Error Back Propagation (EBP) for classification of different eye images. To reduce the complexity of layered neural network the dimensions of input vectors are optimized using Singular Value Decomposition (SVD). The main of this work is to provide for best method for feature extraction and classification. The details of this combined system named as SVD-EBP system, and results thereof are presented in this paper. Keywords- Singular value decomposition(SVD), Error back Propagation(EBP).