A Permutation Group-Based Evolutionary Algorithm for Car Sequencing Problems in Assembly Lines

The car sequencing problem (CSP) in assembly lines has been a significant focus for automobile manufacturers over several decades. Although various optimization methods have been devoted to solving it, most existing work only handled simple CSPs with a few variables and naive objectives, unable to obtain diverse optimal solutions meeting the complicated requirements in real-world scenarios. Therefore, this paper formulates CSPs into a large-scale combinatorial many-objective optimization problem, and solves it by developing a multi-objective evolutionary algorithm based on permutation group. The proposed algorithm suggests permutation group-based crossover and mutation operators to preserve the excellent genes of parents and enhance the search ability. Moreover, by extracting heuristic information from the current population, a probability matrix is established to efficiently repair solutions. To evaluate the performance of the proposed algorithm, empirical comparisons are carried out on nine real-world datasets having up to 1000 variables. The experimental results demonstrate that the proposed algorithm has significant advantages over several state-of-the-art evolutionary algorithms.