CNF and DNF considered harmful for computing prime implicants/implicates

被引:31
|
作者
Ramesh, A
Becker, G
Murray, NV
机构
[1] SUNY ALBANY, INST PROGRAMMING & LOG, DEPT COMP SCI, ALBANY, NY 12222 USA
[2] INTEL CORP, CHANDLER, AZ 85226 USA
基金
美国国家科学基金会;
关键词
solution; prime implicants/implicates;
D O I
10.1023/A:1005721905269
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Several methods to compute the prime implicants and the prime implicates of a negation normal form (NNF) formula are developed and implemented. An algorithm PI is introduced that is an extension to negation normal form of an algorithm given by Jackson and Pais. ir. correctness proof of the PI algorithm is given. The PI algorithm alone is sufficient in a computational sense, However, it can be combined with path dissolution, and it is shown empirically that this is often an advantage. None of these variations rely on conjunct ive normal form or on disjunctive normal form. A class of formulas is described for which reliance on CNF or on DNF results in ar. exponential increase in the time required to compute prime implicants/implicates. The possibility of avoiding this problem with efficient structure preserving clause form translations is examined briefly and appears unfavorable.
引用
收藏
页码:337 / 356
页数:20
相关论文
共 20 条