Solving a Virtual Resequencing Problem in Automotive Paint Shops with Ant Colony Optimization

Abstract

Virtual resequencing can be used to achieve color-batching in automobile paint shops, in which preassigned colors are rearranged to white car bodies while their sequence positions in the upstream body shop remain unchanged. The objective is to minimize the number of color changes during painting. A 0-1 integer programming model is presented to describe the color-batching problem. Aimed at quickly finding near-optimal solutions to this NP-complete problem, an ant colony optimization (ACO) metaheuristic is proposed. Computational experiments show that the ACO algorithm generates as good results as CPLEX for small and medium size instances. For large size instances in real production settings, the algorithm can be applied effectively on a rolling horizon basis to obtain favorable results.