电子学报 ›› 2017, Vol. 45 ›› Issue (7): 1601-1607.DOI: 10.3969/j.issn.0372-2112.2017.07.008

• 学术论文 • 上一篇    下一篇

基于三值多样性粒子群算法的MPRM电路综合优化

俞海珍, 汪鹏君, 张会红, 万凯   

  1. 宁波大学电路与系统研究所, 浙江宁波 315211
  • 收稿日期:2016-05-10 修回日期:2017-04-15 出版日期:2017-07-25
    • 通讯作者:
    • 汪鹏君
    • 作者简介:
    • 俞海珍,女,1975年出生,浙江象山人,博士,高级实验师,主要研究方向为低功耗集成电路理论及优化设计.E-mail:yuhaizhen@nbu.edu.cn;汪鹏君(通信作者),男,1966年出生,浙江奉化人,博士、教授、博士生导师,中国电子学会高级会员,中国计算机学会高级会员,中国电子学会电子线路与系统专业委员会委员,中国计算机学会多值逻辑与模糊逻辑专业委员会委员.主要研究方向为多值逻辑和低功耗集成电路理论及优化设计.E-mail:wangpengjun@nbu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.61306041,No.61234002); 宁波市自然科学基金 (No.2016A610065); 学校科研基金 (No.XKL15D225)

Optimization of MPRM Circuits Based on Ternary Diversity Particle Swarm Optimization

YU Hai-zhen, WANG Peng-jun, ZHANG Hui-hong, WAN Kai   

  1. Institute of Circuits and Systems, University Ningbo, Ningbo, Zhejiang 315211, China
  • Received:2016-05-10 Revised:2017-04-15 Online:2017-07-25 Published:2017-07-25
    • Supported by:
    • National Natural Science Foundation of China (No.61306041, No.61234002); Ningbo Natural Science Foundation of Zhejiang Province (No.2016A610065); Scientific Researxh Fund of Ningbo University (No.XKL15D225)

摘要:

通过对离散三值粒子群算法的研究,提出一种三值多样性粒子群算法以求解MPRM(Mixed-Polarity Reed-Muller,MPRM)电路综合优化问题.首先根据混合极性XNOR/OR展开式的特点和几率换算法则,推导出三值粒子群算法的运动方程,在此基础上,采用广泛学习策略和三值变异操作进行算法改进;然后建立三值多样性粒子群算法的粒子与MPRM电路极性的参数映射关系,结合估计模型和XNOR/OR电路混合极性转换方法,将所提算法应用于MPRM电路的最佳功耗和面积极性搜索;最后对10个PLA格式MCNC Benchmark电路进行测试.结果表明:与已发表的方法相比,该文的优化算法表现出了总体显著性的性能优势.

关键词: 三值多样性粒子群算法, MPRM电路, 逻辑优化, 极性转换

Abstract:

Based upon the research of discrete ternary particle swarm optimization,the ternary diversity particle swarm optimization (TDPSO) is proposed for optimization of mixed polarity reed-muller circuits (MPRM).According to the characteristics of mixed polarity XNOR/OR expression and probabilistic transition rules,motion equation of discrete ternary particle swarm optimization (DTPSO) is deduced.On the basis of DTPSO,a TDPSO algorithm is proposed by introducing extensive learning and ternary mutation.A mathematic mode for area and low power dissipation is built and mixed polarity conversion of XNOR/OR circuits is improved.Then on the basis of parameter mapping relations between particle and polarity,TDPSO is introduced to search the best polarity for power and area of MPRM circuits.Finally,the proposed algorithm is tested on ten circuits from MCNC benchmark circuits,and the experimental results show that the algorithm significantly outperforms the reported method.

Key words: TDPSO algorithm, MPRM circuits, optimization, polarity conversion

中图分类号: