Truncated Sum MVL Minimization Using the Neighborhood Decoupling Algorithm.

摘要

Several heuristics have been developed for the multiple-valued logic minimization problem, and while each claims some advantages in specific examples, none is significantly better than the others. Heuristic methods are interesting because exact minimization methods are extremely time-consuming. With the computer software developed at NPS called HAMLET, users can easily investigate their own heuristics. The primary goal of this thesis is to develop an algorithm that makes the minimization of multiple-valued logic functions reasonably close to the optimal solution. The neighborhood decoupling (ND) algorithm is built on top of HAMLET. The idea of the ND algorithm is: always select the most isolated minterm as well as choose the most isolated implicant. In this thesis, the implementation of the ND algorithm is described. A performance analysis of the ND algorithm is presented by comparing results and computation time with two published algorithms, Pomper and Armstrong's and Dueck and Miller's. (Author) (kr)