Scheduling of Heterogeneous FDM Printer Farms under Sequence-Dependent Setups and Maintenance: A Discrete Random-Key GBFIO (DR-GBFIO) Approach
CMES-COMPUTER MODELING IN ENGINEERING & SCIENCES, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.32604/cmes.2026.091343
- Dergi Adı: CMES-COMPUTER MODELING IN ENGINEERING & SCIENCES
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Compendex, INSPEC, zbMATH, Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Kocaeli Üniversitesi Adresli: Evet
Özet
The scheduling of heterogeneous parallel machines under complex operational constraints is a computationally demanding NP-hard combinatorial optimization problem. While often encountered in modern production farms, resolving sequence-dependent setups and mandatory maintenance windows requires advanced metaheuristic frameworks. This study formulates an integrated scheduling model for joint job-to-machine assignment and sequencing, utilizing a heterogeneous Fused Deposition Modeling (FDM) printer farm as the application domain; the model requires every job to be assigned and completed, and its primary objective is the minimization of the maximum completion time (makespan), with an extended weighted-sum multi-objective formulation also provided. To solve this problem, the recently proposed Grizzly Bear Fat Increase Optimizer (GBFIO) is faithfully adapted to the discrete domain via random-key encoding and an event-driven decoder, yielding the proposed Discrete Random-Key GBFIO (DR-GBFIO). DR-GBFIO is computationally benchmarked against Particle Swarm Optimization (PSO), a Genetic Algorithm (GA), and Harris Hawks Optimization (HHO) across three problem scales reaching up to 20 machines and 100 jobs over 1200 optimization runs, evaluated via Friedman and Wilcoxon signed-rank tests. Results demonstrate that DR-GBFIO significantly outperforms all baselines across all three problem scales, with an approximate 4% to 18% reduction in makespan, a pattern consistent with preserving the original equations, including multi-source attraction, linearly decaying envelopes, and repeated intensification, without approximation. This study contributes a robust metaheuristic framework and one of the first statistically validated, equation-faithful adaptations of GBFIO to a combinatorial context.