基于改进型模糊神经网络的设计方案评价方法

发布时间：2018-06-28 09:26

本文选题：方案评价 + 模糊神经网络　；参考：《组合机床与自动化加工技术》2017年11期

【摘要】：产品概念设计方案的准确评价是产品设计开发成败的重要因素。针对当前产品概念设计方案评价方法的不足,结合模糊神经网络的评价方法,提出了一种基于改进遗传算法的模糊神经网络评价方法。通过建立模糊评价模型完成网络训练,并采用将遗传算法与BP算法相结合的多子群自适应遗传BP算法优化,在全局范围内快速找到最优解。并通过Matalab实现了对迷你型冰箱概念设计方案数据的训练和样本测试,预测结果验证了该方法的准确性和实用性。
[Abstract]:The accurate evaluation of product conceptual design scheme is an important factor for the success or failure of product design and development. In view of the deficiency of the current evaluation method of product conceptual design scheme and the evaluation method of fuzzy neural network, a fuzzy neural network evaluation method based on improved genetic algorithm is proposed. The fuzzy evaluation model is established to complete the network training, and the multi-subpopulation adaptive genetic algorithm is adopted to optimize the network, which combines genetic algorithm with BP algorithm to quickly find the optimal solution in the global range. The training and sample test of conceptual design data of mini refrigerator are realized by Matalab. The prediction results verify the accuracy and practicability of this method.
【作者单位】：四川大学制造科学与工程学院;
【基金】：国家自然科学基金资助项目(51435011) 四川省科技支撑计划资助项目(2014GZ0124)
【分类号】：TP183

【相似文献】