基于模式联想的结构布局优化方法

本文把结构布局优化问题分为两个阶段，即基于模式联想的方案优选和基于数值模型的几何优化和参数优化。提出运用模式联想方法产生出多个结构布局方案然后优选出一个或多个可行性实用的方案，再由基于数值模型的几何优化和参数优化方法获得优化的结构布局。运用人工神经网络技术，通过对大量已成功的结构布局实例的实习，获得具有知识和经验的神经网络系统，然后使用网络可对某个输入集实现模式联想，产生初步布局，进而结合人工智能

Approach for configrational structural optimization based on pattern association

The configrational structural optimization is divided into two phases in this paper.That is ,the scheme optimization based on pattern association and the geometric as well as parametric optimization based on mathematic models,The schemes of structural configration are produced by pattern association,and then one or more practicable schemes are optimized .Pattern association is based on neural network technique.The network can learn the knowledge of structural configration by large number of successful examples collected.After the learning is successful,the network will able to produce several initial schemes as the response to the input sets.The optimization of these schemes is based on artificial intelligence techniques.The key techniques of the approach are discribed in detail.An example is given to show the practicableness of the approach.