Abstract: In order to improve the effectiveness of the bulk cargo port productivity evaluation,a method based on Gauss harmony algorithm（GHS）and the improved functional link neural fuzzy network are combined to evluate bulk cargo port productivity. Firstly, according to the evaluation index of the operating properties, the comprehensive coverage of the model parameters is designed, and the selection of evaluation indicators of productivity bulk cargo port is realized by combining the actual property of data acquisition and process analysis; Secondly, the bulk port productivity evaluation model using the functional link neural network is designed, and the construction of bulk cargo port productivity evaluation model is realized by taking the network as the output of the network for model design of fuzzy rules; Finally, through simulation experiments, the fitting degree between the actual and expected outputs of the model is very close, which can achieve more than 95% sample data recognition efficiency, and can meet the accuracy requirement of bulk cargo port productivity evaluation in real bulk cargo port.

Key words: Harmony search algorithm, functional neural network; bulk cargo port, productivity evaluation

摘要： 为提高散杂货港口生产力评估的有效性，提出一种基于高斯和声算法（GHS）优化功能链接模糊神经网络的散杂货港口生产力评估方法。首先，根据评价指标的可操作属性，对模型指标的覆盖全面性进行设计，并结合实际属性对数据采集处理过程进行分析，实现对散杂货港口生产力评估的指标选取；其次，利用泛函链接神经网络进行散杂货港口生产力评估模型的设计，并将其作为网络输出进行网络模型的模糊规则设计，实现了散杂货港口生产力评估模型的构建；最后，通过仿真实验，该模型的实际与期望输出之间的拟合程度非常接近，可实现95 %以上的样本数据的识别效率，可以满足真实散杂货港口散杂货生产力评估的精度要求。

关键词: 和声搜索算法, 泛函神经网络, 散杂货港, 生产力评估