Boolean neural network realization for mirror symmetry cases

Boolean neural networks are a class of neural networks in which input and output are Boolean. A general algorithm for mirror symmetry cases in Boolean neural network is presented. The new algorithm provides a systematic transform method to reduce the number of compact zones (hidden units) instead of trial method for the mirror symmetry cases with IMS method previously available in literature. The properties of the algorithm are given and demonstrated by some examples. A simulation of the proposed algorithm has also been made.