Mixed Model Assembly Line Sequencing Problem By Considering The Conveyor Stoppage Time

As a branch of the workshop scheduling problem, the efficient use of mixed model assembly line largely depended on the sequencing of mixed model assembly line. A reasonable sequencing scheme can improve the market competitiveness of enterprises by reducing the production cost, product inventory, and make it possible to achieve zero inventory and Just in Time Production. The problem has been recognized as an NP problem, and widely concerned by researchers for decades.This article discussed about the single objective and multi-objective optimization of mixed model assembly line sequencing problem. Different from most researchers who only considered the conveyor stoppage time and ignored the worker’s leisure time when they studied the mixed model assembly line sequencing problem.this article proposed to unify the conveyor stoppage time and worker’s leisure time as one optimization objective by giving them different weights. A single objective optimization model was established by considering the influence of adjustment time on the production line. Genetic algorithm and differential evolution algorithm were used to solve three different sequencing problems based on the established optimization model. Finally, the advantages and disadvantages of the two algorithms were analyzed and compared based on the problems.Based on the single objective optimization model, a multi-objective optimization model was established by focusing on three different objectives: minimization the conveyor stoppage time and worker’s leisure time, minimization the adjustment time and minimization the early/delay delivery time. This article proposed an advanced double differential evolution algorithm by studying the easily losing part(non-dominated individuals shortcoming) of the differential evolution algorithm and double differential evolution algorithm. In order to reduce the abandoned rate of non-dominated individuals, an assistant population was used to save the abandoned non-dominated solution on selection operation, to randomly compare the individuals of assistant population with the individuals of father population and main population, and to select the better individual into the next generation. In order to save the non-dominated individuals as more as possible, this article proposed to directly select the non-dominated individuals of the three populations into the next generation after some iterations. The improved algorithm was then compared with NSGAII, DE and double differential evolution algorithm. All the results showed the superiority of the proposed algorithm.