A feature based neural network model for distributed flexible flow shop scheduling considering worker and transportation factors

Abstract

In the context of distributed flexible flow shop scheduling (DFFSP), the factors of worker and transportation have a significant impact on the production efficiency within a manufacturing environment. However, previous research rarely considers both worker and transportation simultaneously. Therefore, this paper investigates a DFFSP with worker and transportation factors (DFFSP-WT). Considering the characteristic of DFFSP-WT, a neural network-based monarch butterfly optimization (NNMBO) is designed to minimize the objectives of makespan, total cost, and worker fatigue. In the NNMBO, the monarch butterfly optimization (MBO) is employed as the primary optimization operator to determine the job sequence. Furthermore, a feature-based search strategy (FSS), which encompasses six distinct local search operators, is developed to enhance the search capability. Additionally, a feature-based neural network model (FNN) is designed to adaptively select the best FSS. To verify the effectiveness of NNMBO, the simulation experiments are conducted with other state-of-the-art algorithms on test instances, the experimental results demonstrate that the NNMBO is a promising algorithm to solve DFFSP-WT.