Answers are grounded in this paper and cite their sources.
Coordinated maintenance at dry bulk terminals requires balancing maintenance execution with the stability of available operating capacity. This study formulates a bi-objective scheduling model that maximizes maintenance-duration workload and minimizes temporal variation in maintenance-induced capacity occupation under line-conflict and capacity constraints. A multi-objective differential learning algorithm with a hybrid offspring strategy (MODL-HOS), built on a multi-objective evolutionary algorithm based on decomposition (MOEA/D), combines base-guided discrete variation, random perturbation and periodic population response. Validation separates exact small-instance benchmarking from comparative evaluation on larger task tables. Four fixed small instances are evaluated against complete Gurobi Pareto objective sets using 30 search seeds per instance. A separate comparison uses 16 medium and large task tables, five search seeds per table and four algorithms, each with 20,100 candidate evaluations. MODL-HOS retains a measurable small-instance approximation gap. On the larger tables it achieves the lowest aggregate mean empirical normalized inverted generational distance (nIGD), 0.028520, and highest mean empirical hypervolume ratio, 96.04%. Six-test Holm correction supports a hypervolume advantage against all three baselines, while the empirical nIGD advantage is significant only against the non-dominated sorting genetic algorithm II (NSGA-II). The findings support conditional comparative performance; the empirical ratios do not certify optimality, and validation across terminal configurations and with practitioners remains necessary.
No takes yet. Share an insight, caveat, or question.
Xu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: