| 基于神经网络的生产作业计划理论研究 |
| |
| 引用本文： | 黄景平,冯珊,周凯波.基于神经网络的生产作业计划理论研究[J].华中科技大学学报(自然科学版),2000,28(9):4-6. |
| |
| 作者姓名： | 黄景平 冯珊 周凯波 |
| |
| 作者单位： | 华中理工大学系统工程研究所 |
| |
| 基金项目： | 国家自然科学基金重大项目(NNSF79990850). |
| |
| 摘 要： | 研究如何利用前向神经网络(也称多层感知机)安排单件小批量订单生产类型企业的生产作业计划,针对数据难于组织、反传算法易陷入局部极小点和输出结果存在偏差的问题,提出了基于位混合编码的样本组织方式;权值和迁跃点记忆队列的网络训练方法和结果校正方法.仿真结果证明了方法的有效性和基于神经网络研究生产作业计划的可行性.
|
| 关 键 词： | 多层感知机 生产作业计划 反传算法 |
| 文章编号： | 1000-8616(2000)09-0004-03 |
| 修稿时间： | 2000年1月13日 |
A Study of Production Active Planning Based on the Neural Network-the Theoretical Part |
| |
| Huang Jingping Feng Shan Zhou Kaibo Inst. of System Eng.,HUST,Wuhan ,China..A Study of Production Active Planning Based on the Neural Network-the Theoretical Part[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2000,28(9):4-6. |
| |
| Authors: | Huang Jingping Feng Shan Zhou Kaibo Inst of System Eng HUST Wuhan China |
| |
| Institution: | Huang Jingping Feng Shan Zhou Kaibo Inst. of System Eng.,HUST,Wuhan 430074,China. |
| |
| Abstract: | The production active planning with multi layered perceptrons (MLP) is explored. Being difficult to organize the data, BP algorithm usually gets stuck in local minimal and the results deviate from the ideal values. Some improved methods for solving the problem are the following: mixed bit encoding based on sample organization mode, memory queue of weight and transitional point and the correction method. The validity of the methods and the feasibility of the research are proved through the simulation result. |
| |
| Keywords: | Multi Layered perceptrons production active planning BP algorithm |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
