求解FPRM电路极性优化问题的改进多目标粒子群算法

摘要

针对多目标要求下较大规模固定极性Reed-Muller (FPRM)逻辑电路的极性优化问题,提出一种基于改进多目标粒子群算法的求解方法.首先根据延时、面积及功耗的综合要求建立FPRM电路极性优化的多目标决策模型;然后利用外部档案库引导粒子种群进行兼顾全局搜索及局部开发的双重更新,并通过Pareto占优进行粒子优劣性评价,以获取满足延时短、面积小、功耗低的最优极性解集;最后利用MCNC Benchmark电路进行性能测试,并与3种当前较优算法进行对比,验证了文中算法的有效性.%To optimize the multi-objective polarity design of large-scale FPRM circuits,a solution based on improved multi-objective particle swarm optimization (IMOPSO) algorithm is proposed.Firstly,the multi-objective decision model is established according to the delay,area and power of large-scale MPRM circuits.Then,based on the principle of the exploration and exploitation,the particles,representing the circuits' polarities,achieve evolution by means of repository and gain quality evaluation from Pareto analysis,to obtain the Pareto optimal set for delay-area-power trade-off.Finally,the proposed solution is compared with the three currently preferred algorithms on MCNC Benchmark with PLA format,and the results verify the effectiveness of the solution.