Manufacturing System Design Considering Stochastic Product Evolution and Task Recurrence

Abstract

The conventional approaches to manufacturing system design and line balancing have often focused on a single generation of products, thus leading to new design or rebalancing when new products are introduced or different models are produced in the same line. As the life cycles of product models become shorter and shorter, this new product then new system-design approach is becoming increasingly ineffective due to too frequent production interruption. Therefore, effective solutions to system-design problems should consider the evolution of products over multiple generations and models, and this paper presents such new methods. Mixed-integer programming models are developed for (1) designing manufacturing system configurations that are cost effective for product evolution involving uncertainty and (2) maximizing the recurrences of manufacturing tasks on the same machines throughout product evolution. A decomposition-based solution procedure is also developed to reduce computational complexity. These new methods can provide a stable system-design solution enabling quick product launches with less line change-over for new or different products.