As one of the new subjects developed in recent years, artifical neural network has been used widely in the areas of signal treatment, mode detecting, knowledge processing engineering or specialist system and controlling for motion process. In the paper, the author made up artifical neural network computer program by means of utilizing forward multy-layer neural theory and adopting error backward propagating method, and applied artificial neural network to the design of propeller.Through the training of som...

            人工神经网络 (artificialneuralnetwork)是近些年兴起的一门新兴学科 ,在信号处理与模式识别、知识处理工程或专家系统、运动过程控制等领域应用得比较广泛 .笔者运用前向多层神经网络理论 ,采用误差反向传播方法 (BP法 )编制了人工神经网络计算机应用程序 ,将人工神经网络应用到船舶螺旋桨设计中 .通过对部分类型螺旋桨图谱的训练 ,并将其数据与回归分析的数据相比较 ,结果表明 :所应用的理论是可靠的 ,精度是足够的 ,特别是对螺旋桨图谱训练的结果是完全具有实用价值的 .

文摘来源

         This paper aims at the interactivity of Distributed ECDIS (Electronic Chart Display and Information System) and users. The Design of Distributed ECDIS component library adopts the UML modeling method and the principle of component technology. First according to theories of component and distributed system, the layer model of ECDIS component library is built. Then the functions of ECDIS are abstracted. Every function becomes a sub-component. At last the sub-components are divided into layers with the layer m...

            针对分布式电子海图系统开发过程中的系统与用户的交互问题,采用了标准建模语言的建模方法,并结合构件技术的原理,设计出分布式电子海图的构件库.首先根据构件与分布式系统的原理与特点,建立电子海图构件库的层次模型.其次对电子海图系统中的各种应用和功能进行提取,每一个功能与应用形成独立的子构件.最后根据构件库的层次模型将子构件进行详细的层次划分,形成一个完整的功能独立的电子海图构件库.该构件库在实际应用中,对于分布式网络环境下的各种功能需求能够表现出较强的稳定性和适应性,是电子海图系统在网络应用方面的核心部分.

文摘来源

         In this paper,the authors make use of the forward multi-layer neural work theory to apply BP network to predicting and evaluating design of ship structure.Through training 10,000~40,000t bulk cargo carrier,the authors compare the results of simulation with the data obtained by fuzzy synthetic evaluation in literatures.The results show that the theory is correct,precise,and valuable.

            运用前向多层神经网络理论,采用改进BP网络,将人工神经网络应用到船体结构设计的预测与评估中,通过对1万~4万t级散货船训练,并将仿真得到的评估结果与文献中利用模糊评判得到的数据相比较,结果表明:所应用的理论可靠、精度足够。