An optimization framework for worker allocation in aircraft final assembly lines based on simulation alternative modelling and historical data

Abstract

The aircraft final assembly line is a highly labour-dependent manufacturing system, which leads to complexity and uncertainty. Random on-field disturbances mean that worker allocation is in an ever-changing state. Therefore, this article proposes a worker allocation optimization framework with efficiency and accuracy. First, an artificial neural network (ANN)-based model is developed as the alternative to simulation to evaluate worker allocations. Secondly, the non-dominated sorting genetic algorithm-II (NSGA-II)-based algorithm and ANN-based alternative model are integrated to realize the optimization of worker allocation. Finally, an initial population generation method combining historical and randomly generated allocations is proposed to make full use of the domain knowledge in historical data. Based on a real case, the ANN-based alternative model shows a strong advantage over the simulation in terms of efficiency, with competitive accuracy as well. The experimental results also indicate that the proposed initial population generation method effectively improves the performance of the algorithm.