A Three-dimension Stacking Model with Modified Genetic Algorithm

Abstract

The three-dimensional stacking problem (3D-SP) is a challenging task in cold chain warehouse. Different from common three-dimensional packing problem, 3D-SP problem is more operable and can be stacked from all directions of the pallet. Based on this characteristic, we construct our model by considering the utilization rate of pallet space and the stability criterion of goods together as objective function. Further, four constraints are designed, which are placement direction, pallet space and no overlapping. According to the characteristics of the problem, a new improved genetic algorithm is proposed. In specific, the order of goods placement is regarded as individual, and with the consideration of order feature, we designed a more reasonable crossover and mutation operator. Compared with traditional greedy and genetic algorithm, our algorithm outperforms them and proved to be effective on 3D-SP problem.