计算机应用与软件2016，Vol.33Issue(5)：248-251,4.DOI:10.3969/j.issn.1000-386x.2016.05.062
基于粗糙集理论的 PSO-IOIF-Elman 神经网络建模
PSO-IOIF-ELMAN NEURAL NETWORK MODELLING BASED ON ROUGH SET THEORY
摘要
Abstract
In view of the advantages and disadvantages of power amplifier modelling in system-level simulation,this paper proposes a method using simplified particle swarm optimisation (PSO)algorithm to optimise the improved OIF-Elman neural network (PSO-IOIF-Elman) power amplifier behaviour model in combination with rough set theory.Considering different influences of small signal and large signal on the PA in regard to nonlinear characteristic of memory effect,and combing the characteristics of AM-AMand AM-PMmodulation distortion,the model describes the self-feedback coefficient of OIF-Elman neural network to the normalised input and output voltage data.It employs the simplified PSO optimisation algorithm for preventing from falling into local optimal,and uses rough set theory to correct and compensate model’s forecast value for improving the prediction precision.Through Matlab simulation comparison,the training error of the model reduces by 9.53% and the convergence rate improves by 11.31%,therefore verify the validity and reliability of the modelling method.关键词
功放记忆非线性/行为模型/IOIF-Elman 神经网络/简化粒子群算法/粗糙集理论Key words
Nonlinear characteristic of power amplifier memory effect/Behaviour model/IOIF-Elman neural network/Simplified particle swarm optimisation/Rough set theory分类
信息技术与安全科学引用本文复制引用
田娜,南敬昌,高明明..基于粗糙集理论的 PSO-IOIF-Elman 神经网络建模[J].计算机应用与软件,2016,33(5):248-251,4.基金项目
国家自然科学基金项目（61372058）；辽宁省高等学校优秀科技人才支持计划项目（LR2013012）；辽宁工程技术大学研究生科研项目（5B2014032）。 （）
