Application of Clustering Strategy and Parallel Processing to the Large-Scale PCB Drilling Path Planning

• Main Authors: Chiu,Chien, 邱謙
• Other Authors: Yau,Hong-Tzong
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/65000798337396947411

碩士 === 國立中正大學 === 機械工程學系暨研究所 === 102 === This thesis proposes a novel path planning algorithm which integrates clustering strategy and parallelized LKH algorithm and is applied to the application of large-scale printed circuit board (PCB) drilling path planning problem. In the PCB produced process, drilling is one of the important processes. The movement distance of tool will affect the processing time. Generally, repeating pattern arrangement will be taken into consideration to save costs when planning. However, many related literatures in the past did not care this unique PCB information. In this study, the repeating pattern features is used for clustering strategy. A efficient repeating pattern finding method drives from Stringology is integrated into this paper. This method can be used to separate the repeating pattern in a efficient way. By the results of clustering strategy, using the LKH algorithm combined with multi-core processors while the path of each region in parallel computing. It can significant reduce the computation time. Finally, the beginning and end points considered in advance, the path will be linked to form a complete path. Finally, TSPLIB and practical cases are used to validate the path planning algorithm proposed by this study. Experimental results show that the savings in computation time has a very significant effect. On the path length, compared with the best path length difference with only less than 5%.