Solving Many-Objective Car Sequencing Problems on Two-Sided Assembly Lines Using an Adaptive Differential Evolutionary Algorithm

Abstract

The car sequencing problem (CSP) is addressed in this paper. The original environment of the CSP is modified to reflect real practices in the automotive industry by replacing the use of single-sided straight assembly lines with two-sided assembly lines. As a result, the problem becomes more complex caused by many additional constraints to be considered. Six objectives (i.e. many objectives) are optimised simultaneously including minimising the number of colour changes, minimising utility work, minimising total idle time, minimising the total number of ratio constraint violations and minimising total production rate variation. The algorithm namely adaptive multi-objective evolutionary algorithm based on decomposition hybridised with differential evolution algorithm (AMOEA/D-DE) is developed to tackle this problem. The performances in Pareto sense of AMOEA/D-DE are compared with COIN-E, MODE, MODE/D and MOEA/D. The results indicate that AMOEA/D-DE outperforms the others in terms of convergence-related metrics.