부산대학교 도서관

The manufacturing landscape is experiencing a trend towards mass customization of products as customer demands continue to rise. Product life cycles are becoming shorter, while order fluctuations resulting from small lot-sizes increase. Traditional manufacturing lines, designed to produce high quantities of identical products, are ill-equipped to face the new challenges. While the paradigms are shifting, adoption of modern technologies for dynamic and flexible shop floor designs, such as cyber-physical systems and multi-agent systems, remains low. A clear pathway towards reconfiguring legacy systems stepwise is missing, and the new technologies are not yet established enough. We aim to alleviate these problems by introducing a novel, general simulation framework, developed with partners from the industry, for evaluation of manufacturing system designs. We highlight the usage of this framework by applying an online optimization model, utilizing techniques from organic computing, to optimize order and task prioritization in a hypothetical minimal modular smartphone production scenario. Our experiments show that the usage of an independent ob-server/controller structure is suited to solving these problems. The outlined scenario and optimization procedure can be extended to cover real-life manufacturing systems of much greater complexity within the provided framework.