BP神经网络和蚁群算法在防灾产品优化中的应用  

Application of BP neural network and ant colony algorithm in optimization of disaster prevention products

作　　者：夏之秋[1] 踪念品[2] 殷孝雎[2] 赵婷婷[2] 夏君铁[3] 

机构地区：[1]英国莱斯特大学科学与工程学院 [2]沈阳工程学院信息工程学院,沈阳110136 [3]大连大学物理学院,辽宁大连116622

出　　处：《辽宁工程技术大学学报（自然科学版）》2014年第12期1711-1714,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：国家"863"高技术研究发展计划资助项目(2001AA412011);国家自然科学基金资助项目(60174037;50275013);辽宁省教育厅一层次优秀人才资助项目(LR2013060);辽宁省自然科学基金资助项目(2013020123)

摘　　要：针对产品优化设计中对应功能结构选择的多样性以及产品的个性化客户需求,利用蚁群算法结合BP神经网络优化的思想进行了定量求解.通过研究产品需求动态变化的计算机理,给出了产品的可变长基因表达方法;根据产品功能与结构之间的相互关系,给出了产品的相关矩阵;采用BP神经网络方法计算相关矩阵的参数值;并且将结构组合方式看成是蚂蚁走过的路径,采用蚁群算法计算出了最优的可选结构.以防灾产品设计为例,结果分析表明:采用这种方法能够实现最优结构选择的自动化设计,运算结果满足客户需求.Aiming at the diversity of structure selection of the corresponding to function in product optimization design and personalized products needs of customers, the quantitative computational method was proposed by using ant colony algorithm in combination with BP neural network. By studying the calculation mechanism of the product dynamic variation, the method of variable length gene expression of product was given. According to the relationship between product function and structure, the correlation matrix of the product is given by BP neural network to calculate the parameters value of the correlation matrix. The combinations of the structure are taken as ants walked path. The optimal structure can be calculated by ant colony algorithm. Example analysis shows that this method can realize the automation design of the optimal structure choice. Calculation results can meet the demand of customer.

关 键 词：基因表达 BP神经网络 蚁群算法 相关矩阵 优化设计 

分 类 号：TB472[一般工业技术—工业设计]