兵工学报
兵工學報
병공학보
ACTA ARMAMENTARII
2009年
z1期
114-118
,共5页
遗传算法%BP神经网络%核心竞争力%航天二部
遺傳算法%BP神經網絡%覈心競爭力%航天二部
유전산법%BP신경망락%핵심경쟁력%항천이부
genetic algorithm%BP neural network%core competitive power%second research academy of China Aerospace Science and Industry Corp.
结合遗传算法与BP神经网络以航天二部为例,建立了其核心竞争力评价模型,该模型利用遗传算法提高了网络收敛的效率,克服了传统的神经网络训练时间长、易陷入局部极值的缺点,保持算法结果的全局最优.以一个算例从验证的角度说明该模型对于评价航天二部核心竞争力具有可行性以及较高的精度.
結閤遺傳算法與BP神經網絡以航天二部為例,建立瞭其覈心競爭力評價模型,該模型利用遺傳算法提高瞭網絡收斂的效率,剋服瞭傳統的神經網絡訓練時間長、易陷入跼部極值的缺點,保持算法結果的全跼最優.以一箇算例從驗證的角度說明該模型對于評價航天二部覈心競爭力具有可行性以及較高的精度.
결합유전산법여BP신경망락이항천이부위례,건립료기핵심경쟁력평개모형,해모형이용유전산법제고료망락수렴적효솔,극복료전통적신경망락훈련시간장、역함입국부겁치적결점,보지산법결과적전국최우.이일개산례종험증적각도설명해모형대우평개항천이부핵심경쟁력구유가행성이급교고적정도.
Taking the second research academy of China Aerospace Science and Industry Corp. as an example, an evaluation model of core competitive power was established by genetic algorithm and BP neural network. The model can improve the convergence efficiency of neural network, conquer the shortcomings of long training and local optimum, achieve whole optimum of algorithm results. A calculation sample was pulled in to illustrate the feasibility and precision of the model from the verified perspective.
