项目名称: 过程神经网络的智能学习算法研究
项目编号: No.60803065
项目类型: 青年科学基金项目
立项/批准年度: 2009
项目学科: 金属学与金属工艺
项目作者: 张军旗
作者单位: 同济大学
项目金额: 20万元
中文摘要: 过程神经网络自提出以来,其基础理论的研究得到较大发展,并在移动通信、太阳黑子预测及民航飞机发动机故障预测等领域取得了显著的实际效果。然而,目前的学习算法过于简化,使得过程神经网络在学习算法上没有充分利用过程或时间的累积效应,缺乏向车辆负载等过程性更强的实际应用推广的有效学习算法。本课题基于对过程神经网络与粒子群算法的深入研究,将过程神经网络与粒子群算法相结合,相互促进,取长补短,提出了基于粒子群算法的过程神经网络智能算法。通过与粒子群算法的有效融合,解决过程神经网络的带有时间累积算子的复杂学习问题,继而在车辆负载预测领域应用,获得了更好的应用效果,实现了过程神经网络向实际应用更加有效的推广。
中文关键词: 过程神经网络;粒子群学习算法;模式识别
英文摘要: Since the process neural network is proposed, its theory has been developed and applied in many fields, such as mobile communication, sunspot pridiction, engine failure prediction, and so on. However, the learning algorithm of the process neural network is so simplified that the process neural network can not fully utilize the accumulative effect of the process or time on the problems like vehicle load. This project proposed the intelligent learning algorithm of the process neural network based on the study and combination of the process neural network and the particle swarm optimization. The proposed intelligent learning algorithm contributes better effects on the real applications to the process neural networks.
英文关键词: Process neural network;Particle swarm optimization;pattern recognization.