论文部分内容阅读
提出一种基于人工神经网络的综合指标评估方法．文章首先提出了一种新的归一化效用函数，把不同类型，不同量纲的原始评估值转换到［－１，１］区间，该效用函数较好地体现了“奖优罚劣”原则，同时对于神经网络又更容易学习和训练；其次，分析了一般多指标评估中各权值的确定方法及存在的困难；第三，详细介绍了基于人工神经网络的多指标综合评价原理及实现方法，并将之实际应用到一柴油机行业的综合经济效益评估中，取得了较满意的结果．文末讨论了基于神经网络的综合评价方法的优点
A comprehensive index evaluation method based on artificial neural network is proposed. The paper first proposes a new normalized utility function to convert the original evaluation values of different types and different dimensions into the [-1,1] interval. The utility function better reflects the principle of “rewarding and punishing inferiorities and penalties” , And at the same time, it is easier to learn and train for the neural network. Secondly, it analyzes the methods of determining the weights and their existing difficulties in the general multi-index evaluation. Thirdly, the principle and implementation of the multi-index comprehensive evaluation based on the artificial neural network Method, and the practical application of a comprehensive evaluation of the economic benefits of the diesel engine industry, and achieved more satisfactory results. The article discusses the advantages of a comprehensive evaluation method based on neural network