Artificial Neural Network Optimization using Layer Partition and Genetic Algorithm along with Statistical Analysis

Two methods of optimizing the feed-forward Artificial Neural Network (ANN) through Genetic Algorithm (GA) are presented. The connection weights of ANN are extracted out. The functions formulated having connection weights as variables are optimized through GA. The optimized connection weights replace the old ANN connection weights. The first method proposed referred to as standard method (ANN-Std) produced closer results. The second method of ANN optimization referred to as ANN-New produced superior results. The method is proved by approximating the diameter of the spindle of cylindrical grinding machine. The second approach (ANN-New) is tested for different number of hidden layer neurons keeping number of layers fixed, the statistical analysis of which is discussed in detail.