A hybrid metaheuristic for two-stage integrated scheduling based on Bayesian classification in disassembly flexible job shops with automated guided vehicles
The uncertainty of structural defects in end-of-life products disrupts disassembly planning and logistics scheduling. This article proposes a two-stage integrated scheduling strategy incorporating Bayesian-classified structural uncertainties. A beta–binomial Bayesian classification framework quantifies the posterior probabilities of structural damage and loss, and a minimum Bayes risk rule classifies each component. A two-stage mixed-integer linear programming model then minimizes the overall makespan covering disassembly job and automated guided vehicle scheduling, solved by an adaptive hybrid genetic discrete grey wolf optimization (AHG-DGWO) algorithm with posterior-guided operators. Experiments on 15 instances from a real power-battery disassembly plant show average makespan reductions of 36.7% and 12.6% over random and independent strategies, respectively; AHG-DGWO achieves the best mean value in 12 out of 15 instances against four metaheuristics, with statistically significant improvements confirmed by Wilcoxon tests and ablation and benchmark studies.
Authors
- Yinghui Ren (ORCID: https://orcid.org/0000-0002-4841-2544)
- Hui Xie (ORCID: https://orcid.org/0000-0003-0674-9410)
- Zengyu Kuang
- Xiaojun Zhuo
- Minghao Shi
Institutions
- Hunan University (CN)
- Changsha Mining and Metallurgy Research Institute (China) (CN)
Publication Details
- Journal
- Engineering Optimization
- Published
- 2026-11-18
- DOI
- https://doi.org/10.1080/0305215x.2026.2755738
- Primary Topic
- Scheduling and Optimization Algorithms
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Hunan Provincial Science and Technology Department
- Science and Technology Program of Hunan Province