或符合全展开式的分解转换算法
Decomposition Conversion Methods for Canonical or-Coincidence Expansions
-
摘要: 针对现有算法在处理多变量实际电路时间开销较大的问题,提出或符合全展开式的分解转换算法.cj最大项展开式和dj最大项展开式在相同极性下两者之间存在转换矩阵,而矩阵运算复杂度较高,把矩阵的运算简化成与和非的位运算,从而大量地节省了运算时间;在此基础上,将cj最大项展开式分解到不同的分组中,提出了分解算法,避免了矩阵的重复计算,再次缩短了计算时间.为了避免cj最大项展开式中过多最大项而造成转化时间开销增加,还提出了基于cj最小项的分解算法.实验结果表明,包含算法适用于处理小变量,但在处理多变量时时间开销增大,而采用了分解算法后,可极大减少转换时间开销.Abstract: Significant computation time is consumed to obtain a better solution for large functions.Decomposition conversion methods are proposed for canonical or-coincidence expansions.Transformation matrix is obtained between cj maxterm and dj maxterm expansions for canonical or-coincidence expansions.An efficient "inclusion" method is proposed using "AND" and "NOT" bitwise operations to simplify matrix operations.As a result,considerable reduction in time is achieved.Based on the "inclusion" methodology,on-set cj maxterms are decomposed into different groups to achieve further time reduction by reusing duplication calculation in matrix operation.cj maxterm expansion and corresponding minterm expansion are used as inputs respectively.Experimental results show that the proposed decomposition methods achieve good performance for not only small variables but large variables as well in terms of conversion time.
下载: