基于神经网络的生产作业计划应用研究

研究利用前向神经网络 (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 ,提出了以下的解决方法 :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 .仿真结果证明了解决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 .

Application of Production Active Planning Based on Neural Networks

e: The production activity planning with multi layered perceptrons (MLP) is explored. In order to solve the problems of data being difficult to organize, BP algorithm usually get stuck in local minimal and the result deviate the ideal values, we advance the methods as follows. They are mixed bit encoding based an sample organization mode, memory queue of weight and transitional point and correction method. The validity of these methods and the feasibilty of the research are proved by the simulation result.