Restructuring multi-level networks by using function approximations

This paper presents a technique for multi-level minimization using implicit don’t cares. One of the major problems in minimization is the calculation of the don’t care set. Conventional approaches project the global don’t cares onto the local support of each node to keep minimization cost low. This prevents restructuring multi-level networks. The approach presented in this paper enables the extension of the support for the projection of the don’t care set, while keeping a low computational complexity. The extension of the support is based on the idea of function approximations. The approach has shown to produce tangible improvements in multi-level networks when compared with existing techniques.