论文部分内容阅读
基于遗传算法和神经网络的多层感知器模型的有机结合,提出一种优化换热器网络的新算法和一种新的编码方法——基因矩阵. 这种算法根据遗传适应度(目标函数)的大小,以随机搜索方式寻找在求解区域的最优解,采用神经网络多层感知器模型实现换热器网络的结构优化和参数优化. 经过遗传-感知模型优化并与外逼近算法做了比较,表明采用此法优化多维、多峰、非凸的换热器网络也具有很好的适应性.
Based on the organic combination of genetic algorithm and neural network model of multilayer perceptrons, a new algorithm to optimize the heat exchanger network and a new encoding method-gene matrix are proposed based on genetic fitness (objective function) , The optimal solution in the solution area is searched by random search method, and the neural network multilayer perceptron model is used to optimize the structure and parameter optimization of the heat exchanger network.After the genetic-perceptual model is optimized and compared with the external approximation algorithm , Which shows that this method to optimize multi-dimensional, multi-peak, non-convex heat exchanger network also has good adaptability.