Research and optimization of the three-dimensional container packing problem for dangerous goods in special-purpose ship holds

The rational packing of dangerous goods in special-purpose ship holds significantly impacts the efficiency of sorting and outbound processes. Under both general constraints of the three-dimensional bin packing problem and the specific constraints of special-purpose ship holds, a Mixed-Integer Programming（MIP） model is constructed with the objective of maximizing the quantity of loaded dangerous goods. A Particle Swarm Optimization Hybrid Genetic Algorithm（PSOGA） is employed, incorporating heuristic rules and average dimension information to effectively accelerate the algorithm's execution. A diversity control mechanism is introduced, along with a two-level search strategy（probability level+time level）, to further enhance search efficiency and solution quality. Packing experiments involving three and five types of dangerous goods were conducted, demonstrating that the algorithm can efficiently solve all instances within 360 seconds, providing a reliable reference for the loading of dangerous goods in special-purpose ship holds.