• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Computational Collective Intelligence >Using Neural Networks as Surrogate Models in Differential Evolution Optimization of Truss Structures
【24h】

Using Neural Networks as Surrogate Models in Differential Evolution Optimization of Truss Structures

机译：用神经网络作为桁架结构差分演化优化的代理模型

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In this study, Differential Evolution, a powerful metaheuristic algorithm, is employed to optimize the weight of truss structures. One of the major challenges of all metaheuristic algorithms is time-consuming where a large number of structural analyses are required. To deal with this problem, neural networks are used to quickly evaluate the response of the structures. Firstly, a number of data points are collected from a parametric finite element analysis, then the obtained datasets are used to train neural network models. Secondly, the trained models are utilized to predict the behavior of truss structures in the constraint handling step of the optimization procedure. Neural network models are developed using Python because this language supports many useful machine learning libraries such as scikit-learn, tensorflow, keras. Two well-known benchmark problems are optimized using the proposed approach to demonstrate its effectiveness. The results show that using neural networks helps to greatly reduce the computation time.
机译：在本研究中，采用强大的成群质算法，利用差分演进来优化桁架结构的重量。所有地图算法的主要挑战之一是耗时的，其中需要大量结构分析。为了解决这个问题，使用神经网络来快速评估结构的响应。首先，从参数有限元分析中收集许多数据点，然后所获得的数据集用于培训神经网络模型。其次，培训的模型用于预测优化过程的约束处理步骤中桁架结构的行为。使用Python开发了神经网络模型，因为这种语言支持许多有用的机器学习库，如Scikit-Learn，Tensorflow，Keras。使用所提出的方法来展示其有效性，优化了两个众所周知的基准问题。结果表明，使用神经网络有助于大大减少计算时间。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号