摘要
复杂对象系统多目标综合评价是一个典型的定性与定量相结合的模糊决策问题。本文将神经网络理论应用于复杂对象系统的综合评价，提出了一种基于Ｂ─Ｐ神经网络的复杂对象系统多目标综合评价方法，并对神经网络的结构、神经网络输入指标属性值的量化方法、神经网络学习及其应用神经网络综合评价的计算机实现算法做了讨论。本文最后介绍了该方法在城市发展水平综合评价中的应用实例。
Multiobjective comprehensive evaluation of complex object systems is a typical fuzzy decision problem with the combination of qualitative and quantitative analysis. By applying the theory of neural network into the comprehensive evaluation of complex object systems. this paper proposes a multiobjective comprehensive evaluation method of complex object systems based on the B-P neural network, describes the structure of the B-P neural network.presents the normalization methods of attribute value of input indexes of the neural network, describes the algorithm of neural network learning and the implementation algorithm in computers of comprehensive evaluation using the neural network.Finally.the application example of this method in comprehensive evaluation of city development level is given.
出处
《管理工程学报》
CSSCI
1995年第1期26-33,共8页
Journal of Industrial Engineering and Engineering Management
基金
中国科学院管理
决策与信息系统开放研究实验室基金
湖北省自然科学基金
关键词
经济
复杂对象系统
多目标
综合评价
神经网络
Complex object system, Multiobjective comprehensive evaluation, Neural networks, B-P algorithm, Self-learning, Membership function, City development level
