Implementation of an Irregular Packaging Algorithm for Multi-Container

Abstract: This study introduces an innovative approach to address a complex, real-world packing problem experienced at Emerson Rosemount Tank Radar. The objectives of the study are manifold: primarily, the optimization of packing efficiency within specified geometric and box constraints, and secondarily, the fulfillment of specific packing objectives. The proposed approach comprises the creation of a distinct placement policy, termed the "Constraint Placement Policy". This policy integrates elements of greedy, bottom-left, and proximity-based alignment heuristics to proficiently guide the packing process. Furthermore, to augment the search for the most efficient packing configurations, a genetic algorithm is designed. The genetic algorithm leverages a three-layered chromosome representation for item sequence, rotation sequence, and box sequence, enabling a flexible and dynamic optimization process. Additionally, this study provides a comprehensive analysis of various metaheuristics, highlighting their relevance and applicability in resolving packing problems. It also devises an automated method to extract model code indices, thereby aiding the creation of a component database for Emerson's products. Thereafter, the packing problem is formulated, considering geometric constraints, box constraints, and packing objectives, which significantly affect the overall efficiency and safety of the packing process.The performance of the packaging algorithm is critically evaluated against shipping data. This comparison accentuates the efficacy of the custom heuristic placement policy and the genetic algorithm, particularly focusing on the genetic algorithm capacity to adapt and optimize packing configurations. A contrast between the proposed hybrid approach and conventional packing methods is also provided, illuminating the strengths and limitations of the proposed solution. The results of this study have considerable implications for Emerson's packing process. Not only does it enhance current packing efficiency, but it also proffers valuable insights for potential future improvements in the industrial packing domain.