基于二进制粒子群优化的卫星典型件工艺知识挖掘  

Binary Particle Swarm Optimization Based Process Knowledge Mining for Typical Parts of Satellite

作　　者：王琳[1] 张永健 钟诗胜[1] 刘金山[3] 

机构地区：[1]哈尔滨工业大学机电工程学院,黑龙江哈尔滨150001 [2]哈尔滨工业大学(威海)船舶与海洋工程学院,山东威海264209 [3]中国空间技术研究院北京卫星制造厂,北京100094

出　　处：《东北大学学报（自然科学版）》2015年第1期119-123,共5页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(51075083);高等学校博士学科点专项科研基金资助项目(20112302130003)

摘　　要：针对卫星典型件在工艺设计过程中设计任务量大、重复性工作多,且其历史工艺数据未能充分有效利用的问题,进行了工艺知识挖掘的研究,以提高工艺知识的重用性.首先对工艺知识挖掘问题进行了描述,建立了工艺知识的关联规则模型;然后针对海量数据中Apriori算法挖掘效率低的问题引入二进制粒子群优化(BPSO)算法,并构造了基于BPSO的关联规则挖掘算法.最后对卫星结构板这一典型件的历史工艺数据进行挖掘,得到了卫星结构板典型工序序列.基于BPSO的关联规则挖掘算法可以有效提高工艺知识的挖掘效率.The huge quantity of design of manufacturing process of satellite typical parts and a lot of repeated jobs existed in the process. Many kinds of process knowledge without reused effectively contained in the historical process data. Process knowledge mining algorithm was studied in order to increase efficiency. The problem was described firstly, and the association rule model was built. In order to improve computational efficiency of Apriori algorithm for huge datasets, binary particle swarm optimization（BPSO） was introduced. Meanwhile association rule mining algorithm based on BPSO was designed. Finally, the designed algorithm was used in process knowledge mining for satellite plate. The mining efficiency of process knowledge can be improved effectively by the association rule mining algorithm based on BPSO.

关 键 词：粒子群优化 典型件 工艺知识 关联规则挖掘 典型工序序列 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]