多维关联函数建模及其低碳实例检索应用  

Modeling of Multi-dimensional Correlation Function and Its Applications for Low Carbon Case Retrieval

作　　者：洪欢欢[1,2] 赵燕伟[1] 陈尉刚[1] 

机构地区：[1]浙江工业大学特种设备制造与先进加工技术教育部/浙江省重点实验室,杭州310014 [2]宁波大学高等技术研究院,宁波315211

出　　处：《中国机械工程》2017年第6期688-694,共7页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(51605231;51275477)

摘　　要：针对低碳设计多维度实例检索建模及其降维计算问题,提出了一种多维关联函数建模及其检索应用方法。首先,通过变斜率原理改进侧距,构建了基于改进侧距的多维侧距模型。然后,针对侧距难以准确描述设计参数与最优值之间距离的问题,建立了多维关联函数及其三层次的低碳实例检索方法。最后以螺杆空压机为例进行了验证,结果表明了改进侧距的正确性和有效性、多维关联函数降维计算的快速性与有效性,以及低碳实例检索的合理性与实例间的差异性。To solve problems of multidimensional case retrieval modeling and its dimensionality reduction calculation for low carbon design,a modeling and application method of multi-dimensional correlation function was proposed.Firstly,an improved side distance method was developed through the principles of variable slope,which provided the basis for developing a multi-dimensional side distance model.Secondly,because side distance formula could not give out the accurate distance values between design parameters and its optimal,a multi-dimensional correlation function which had good dimension reduction effect was established.Then a retrieval method was given based on the multi-dimensional correlation function.Finally,the proposed method and algorithms were verified in the design of a screw air compressor.The evaluation results show that the improved side distance is correct and effective,the dimension reduction method for the multi-dimensional correlation function is fast and viable,and the otherness among these cases is very obvious.

关 键 词：多维关联函数 低碳实例 相似检索 空间降维 螺杆空气压缩机 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]