Collaborative Optimization Method For Batching And Packing In Customized Production Of Plate Products

Plate product is made with the plate as the main raw materials,characterized by plane processing as the main manufacturing method,and formed from plate parts with lightweight assembly relations.It is the common component of PCB,plate furniture,3C products,glass products,sheet metal,building materials,printing industry,and many other products.Plate product manufacturers usually face with the customized demands on the multi-variety and small batch production,these characteristics of the production orders change frequently.The delivery time is tight.It leads to frequent the changeover of the production processes and massive production batch,which causes a low utilization rate of raw materials,low production efficiency,and low equipment utilization rate.A series of problems,such as unreasonable resource allocation,difficult delivery and cost control,urgently call for improving the ability of order batching and packing,and reducing the production batch.Stock packing is usually the first working procedure in the manufacturing process of plate product,it directly affects the composition of production batch and the subsequent technological process,the collaborative optimization of order batching and packing is benefit to achieve systemic optimization in customization production process of plate product,improve utilization rate of raw materials and reduce the production batch,and lower the production cost and improve production efficiency.The study of collaborative optimization problem about order batching and packing has an important theoretical research value and practical engineering significance.The dissertation focuses on the problem of packing and order batching optimization,and the collaborative optimization method of batching and packing in the customized production of plate products.The specific research works are as follows:(1)Research were done for the common demand analysis and optimization problems in the customized production of plate products.The custom design requirements and optimization problems in the production of three types of plate products,such as plate furniture,hollow glass,and PCB sample board,were analyzed in-depth,and the similarity transferred process(geometry-process-motion-optimization)of the production characteristics of the plate products were studied.It reveals that this type of product has the common coupling optimization problem "batch-packing-X."A solution method and idea of subproblem solving algorithm and decoupling algorithm were proposed.(2)A Block-Based Improved Heuristic Search Algorithm(BBHSA)was proposed to the optimization packing problems of two-dimensional rectangle with variable length and guillotine cut constraints.In BBHSA,rectangular pieces were combined into blocks;these blocks provided good layout elements and served as a basic part of the construction tree in the heuristic search process.In order to speed up the search process and improve the quality of the solution,three operations were used in the packing search process:placing and splitting,relaxing and shrinking,and online grouping.The algorithm has good computational efficiency for the case of zero-waste on the benchmark,and it can get optimal solutions for almost all the zero-waste benchmark cases reported in the literature.(3)A multi-fork tree recursive search algorithm based on stacking number layering was proposed for the collaborative optimization of cutting and packing.A mathematical model for minimizing the total cost caused by raw materials and cutting processing time was established.A recursive search method of packing based on stacking number stratification was proposed.A hierarchical iterative collaborative optimization mechanism based on a multi-fork tree structure was adopted.Multi-fork tree hierarchical iteration was performed by using multiple utilization thresholds.The comparison with the mainstream software shows that this method can improve the utilization rate of raw materials and cutting efficiency,and reduce the production cost.(4)A surrogate model of predicting stock packing rate was proposed on the basis of machine learning,aiming at the problem of large calculation time for large-scale packing optimization.We combine the Random Forest model,the XGBboost model,and the Lightgbm model to analyze and learn the historical order packing experience data.The results of the optimization of the new order task can be accurately predicted.(5)An iterative optimization method of order batching and packing based on the stock cutting rate surrogate model was proposed,for the collaborative optimization of order batching and packing in the customized production of plate products.A cohesive hierarchical clustering algorithm that satisfies the delivery time and production process constraints was proposed to optimize the order batching optimization.The surrogate model were used to predict the packing utilization rate and screen the evaluation of the plan,the plan of order batching was screened by the surrogate model before the iterative optimization of order batching and packing problems,which can greatly reduce the search time.Based on the collaborative optimization methods of order batching and packing,the related system was developed and applied in the enterprise,and the value in the engineering practicability of these algorithms was verified.These methods and algorithms can provide efficient services in order batching and packing optimization...