Similarity measures for PC board grouping
Conference Paper
Overview
Additional Document Info
View All
Overview
abstract
In order to cluster (or group) printed Circuit (PC) boards, numerical similarity measurements are required to mathematically quantify relationships among them. In the case of clustering PC boards for the purpose of assembly, variables that affect the production throughput and machine setup time are considered. These variables have been identified as component type, component frequency, and assembly time. The PC boards should be grouped by the similarity that is exhibited between boards with respect to these parameters. However, only component similarity has been used to group PC boards into families. A Placement Location Metric (PLM) methodology that computes the geometric association (Euclidean distance between components and frequency of components) between two boards was developed and tested, providing a high correlation to simulated geometric measures.
