A surrogate-assisted heuristic approach for the joint optimization of resource allocation and scheduling of an aircraft final assembly line

Abstract

In an aircraft final assembly line, there exists flexibility in resource allocation in the sense that processing time of a task can be reduced by allocating more workers to the task. As task processing times change, assembly task scheduling needs to be adjusted accordingly, which in turn has an impact on resource allocation. This paper studies the joint optimization problem of resource allocation and task scheduling. Both resource requirement and starting time of each task need to be determined such that the total resource investment cost is minimized. We propose a surrogate-assisted heuristic approach to solve the problem by decomposing it into two phases. In the first phase, a multi-start iterative search algorithm is developed to perform a dedicated exploration on resource allocation. A surrogate model trained on numerical data is applied to evaluate the resource allocation strategy. In the second phase, a hybrid genetic algorithm embedded with a novel local search procedure is designed for task scheduling. Meanwhile, a fine search for resource allocation can also be achieved. Computational experiments demonstrate the superiority of our proposed approach and that the surrogate plays a significant role in terms of improving solution quality. A realistic case study provides valuable managerial insights for real production.