摘要
为缩短布尔函数系统混合极性Reed-Muller(mixed-polarity Reed-Muller,MPRM)电路面积优化过程的时间,提出了能在任意极性值的MPRM间进行极性转换的系数矩阵变换方法.使用系数矩阵表示布尔函数系统,通过对系数矩阵进行分隔,使用置换和折叠操作完成MPRM极性转换以加快极性转换速度;在此基础上,给出了适用于较大规模MPRM电路的面积优化算法,其中使用遗传算法进行极性空间搜索,并采用基于最短个体距离的适应度计算方法进一步缩短优化过程中的极性转换时间.实验结果表明,与其他MPRM极性转换方法相比,文中方法能够提高MPRM电路面积优化的速度.
In order to reduce the time consumed by area optimization of mixed-polarity Reed-Muller (MPRM) circuits for Boolean function system, a coefficient matrix transformation method for polarity conversion between MPRMs with any polarity number is proposed in the paper. The proposed method uses coefficient matrix to represent Boolean function system, and accelerates polarity conversion by using permutation and folding operations after separation of the coefficient matrix. Based on the proposed method, an area optimization algorithm for large scale MPRM circuits is presented, which takes genetic algorithm as search method for exploring polarity space and employs shortest distance between individuals based fitness calculation for further reducing the time overhead of polarity conversion during the process of area optimization. Experimental results show that, in comparison with other polarity conversion methods, the proposed method can improve the speed of area optimization process for MPRM circuits.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2013年第1期126-135,共10页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(60903033)