摘要
针对传统预测方法无法综合分析多维参数中存在的空间聚合及时间累积效应的问题,该文利用离散过程神经网络对装备技术状态多维参数进行预测。针对网络训练中存在的易获得局部最优解的问题,利用混沌粒子群算法对网络学习过程进行了优化。在此基础上,以某装备传动箱油液数据预测为例对该预测方法的有效性进行了验证,优于其他同类预测方法。
Conventional forecasting methods cannot systematically analyze the aggregation of space and time in multidimensional parameter analysis. To solve the problem, a prediction method based on discrete process neural networks is proposed in this paper. In order to avoid choosing a local optimal solution during the training of the net, the chaotic particle swarm optimization algorithm is introduced in the process of training. Finally, a case study is presented to illustrate the validity of the proposed method.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2016年第6期923-928,共6页
Journal of University of Electronic Science and Technology of China
基金
部级基金
关键词
混沌粒子群算法
技术状态预测
离散过程神经网络
空间聚合
chaotic particle swarm optimization algorithm
condition prediction
discrete process neural networks
space aggregation