Title :
A Parallel Genetic Algorithm for Optimizing an Industrial Inspection System
Author :
Gonzalez Bulnes, Francisco ; Usamentiaga, Ruben ; Fernando Garcia, Daniel ; Molleda, Julio
Author_Institution :
Univ. de Oviedo, Gijon, Spain
Abstract :
Periodical defect detection is a task of great importance during the production of web materials. It can reduce the appearance of a large number of surface defects, which is of vital importance to keep the product quality. In this article, a system used to detect these defects is optimized. This is carried out by looking for the optimal values for each of its configuration parameters. Since the search space formed by these parameters is very large, it cannot be explored exhaustively. For this reason, an intelligent search, like genetic algorithms, must be used. Because the fitness function is computationally heavy, a single computer would take a long time to provide an acceptable solution. For this reason, a cluster of computers is used instead, running a parallel genetic algorithm. Thus, the optimal configuration could be determined in only a few hours.
Keywords :
genetic algorithms; inspection; parallel machines; product quality; production engineering computing; configuration parameters; industrial inspection system; intelligent search; parallel genetic algorithm; periodical defect detection; product quality; surface defects; web materials production; Computational modeling; Computers; Genetic algorithms; Inspection; Manuals; Silicon; genetic algorithm; inspection system; parallel; periodical defects;
Journal_Title :
Latin America Transactions, IEEE (Revista IEEE America Latina)
DOI :
10.1109/TLA.2013.6710381