A hybrid model for mix-bank buffer content determination in automobile industry

Abstract

In mixed-model automobile assembly lines, paint defective vehicles are the main reason for unintentional sequence alteration that stirs up the production sequence so that resequencing is required. This study aims to decide optimal spare vehicles that are held in the mix-bank buffer to be replaced instead of defective vehicles in order to regain the production sequence. We develop a hybrid solution methodology in which optimal spare vehicle content is determined by genetic algorithm (GA) and the releasing order of the vehicles to final assembly (FA) are decided by stochastic mixed integer programming (MIP) model. In addition to discussing the efficiency of the hybrid model, the following insights were gained: 1) not only FA but also paint shop constraints should be considered in production sequence determination when mix-bank buffer is efficiently used; 2) the decrease in defect rate improves sequence restoration linearly; 3) effect of an additional lane on SSAR increase is diminishing.