To kit or not to kit: optimizing part feeding in the automotive assembly industry

Abstract

This is a summary of the author’s PhD thesis supervised by Hendrik Van Landeghem and defended on 22 December 2011 at the Universiteit Gent. The thesis is written in English and is available from the author upon request at veronique.limere@ugent.be and from https://biblio.ugent.be/publication/1985447. This work deals with the materials supply of automotive assembly lines. Mathematical optimization models are proposed for the tactical choice between two opposite materials supply systems, i.e. line stocking and kitting, and a theoretical basis for hybrid materials feeding policies is developed. The automotive industry is characterized by mass customization. Customers are very demanding and want products tailored to their needs. As a consequence, production in the automotive industry takes place on mixed-model assembly lines and a lot of different variant parts need to be supplied to the line. Because of the high number of transactions, efficient material handling is of great importance. An additional distinguishing feature of the automotive industry is the diversity of parts being assembled at the line, ranging from small (e.g. bolts and nuts) to large and voluminous (e.g. wheel well liners and truck side mirrors). This diversity of parts suggests that different methods of line feeding might be appropriate. In industry, different materials supply systems are found. Two opposites in the spectrum of materials supply systems are line stocking and kitting. Both systems have been studied separately, but no general purpose model that would facilitate the selection process—between kitting, partial kitting, and line stocking (i.e. no kitting)—has been reported in literature. This is the topic of this PhD study. A first mixed integer linear programming model (MILP) is developed to assign individual parts to one of both materials supply system alternatives to minimize total costs. The costs considered are the average yearly labor costs for operator picking at the line, internal transport, the kit assembly operation and replenishment of the supermarket. Based on some initial results, a second improved model is obtained taking into