Citation

Pang S, Chen MC. Comput. Ind. Eng. 2023; 180: e109218.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.cie.2023.109218

PMID

unavailable

Abstract

With crew scheduling and Artificial Intelligence (AI) gaining attention in the past few years, there has been growing interest in algorithms that could effectively handle the crew scheduling problem (CSP) in railway transport. Despite AI becoming pervasive in most engineering domains, there is a lack of methods that can provide high-quality crew scheduling in the railway industry. To fill this gap, this study designs a railway crew scheduling model of mixed integer linear programming (MILP) problem utilizing the bacterial foraging algorithm (BFA) and evaluates the model's advantages and limitations. BFA is a novel class of biologically inspired stochastic global search methodology that is based on E. coli bacteria's foraging behavior. Using the Taiwan Railways dataset, we compare the performance of the proposed BFA-based method for railway crew scheduling optimization (BFARCSO) against two benchmark methods, the genetic algorithm (GA) and the particle swarm optimization algorithm (PSO) and showcase its advantages in terms of solution quality and computation time. Finally, a series of computational testing and validation highlights the efficiency and superiority of BFARCSO. It demonstrates that this approach significantly improves the large-scale railway crew scheduling problem over typical approaches, making it well-suited to practice real-time decision support in railway crew scheduling.

Language: en

Keywords

Artificial intelligence; Bacterial foraging algorithm; Crew scheduling; Mixed integer linear programming; Optimization; Railway transportation