【摘要】 本工作通过选取适当的拓朴结构,特别是对零电阻转变温度 T_c 的数据采用压缩与还原的技术处理,建立了一个基于误差反向传播(简记为 B-P)算法的人工神经网络。该网络可根据制备含 F 铋系高温超导材料的几个主要实验参数,相当精确地预言此类超导体的 T_c。本方法显然具有一般意义,从而有力地促进了材料的计算机辅助设计的研究工作,使之进一步趋于成熟。更多还原

【Abstract】 By selecting a proper topologie structure,especially by compressing and re- ducing the Tcdata,we have established an ANN(artificial neural network)Based on the B-P learning algorithm,which can be used to predict guite accurately Tc of the F-droped ceramic supercondoctors from some of their experimental parameters.The present work is of gener- ality and thus has prompted the development of computer-aided materials design.更多还原