论文部分内容阅读
BP网络用于梯度功能材料(EGM)制备过程中的材料特性估计,在解决其制备知识自学习方面,取得了较好的效果。为了进一步改善网络的泛化特性,提高估计精度,本文采用一种改进的双BP算法,并在梯度功能材料合成设计中进行了应用,明显地提高了估计精度。
The BP network is used to estimate the material properties during the preparation of gradient functional materials (EGMs), and has achieved good results in solving the self-learning of preparation knowledge. In order to further improve the generalized characteristics of the network and improve the estimation accuracy, an improved double-BP algorithm is used in this paper. It is applied in the composite design of gradient functional materials, which obviously improves the estimation accuracy.