Uncertain Analysis Based on Milk-Runs Systems Using Bayesian Networks

Abstract

For several years, the role of original equipment manufacturers (OEM) has been changing. Now, their primary operations focus on design, assembly systems, and marketing. Considering that component suppliers provide more than 95% parts of a light vehicle, the complexity of the logistics is high. Seeking to comply with the Just in Time (JIT) philosophy and thus avoid the shortage of components on assembly lines, the companies use the inbound logistics to perform these operations from suppliers to OEMs. The Milk-run concept refers to carrying out the consolidated supply of components to OEMs through the planning of routes considering xed and cyclical time windows. Currently, Milk-run (MR) systems suffer from interruptions in component deliveries; this can produce monetary losses per minute of delay in the assembly line of OEM. Problems in MR have been approached from the perspective of optimization models with variations within them, and there are a few studies about variability and uncertainty in MR operations. The results of this research are favorable because not only uncertainty is quantified; the causes of optimistic and pessimistic scenarios are also quantified.