In this paper, an effective method for identifying and realizing linearly separable Boolean functions (LSBF) of six variables via Cellular Neural Networks (CNN) is presented. We characterized the basic relations between CNN genes and the truth table of Boolean functions. In order to implement LSBF independently, a directed graph is employed to sort the offset levels according to the truth table. Because any linearly separable Boolean gene (LSBG) can be derived separately, our method will be more practical than former schemes [Chen & Chen, 2005a, 2005b; Chen & He, 2006].
