CN-121235360-B - Combined optimization method and system for scheduling production, batch and sample arrangement of plate-type products

CN121235360BCN 121235360 BCN121235360 BCN 121235360BCN-121235360-B

Abstract

The invention relates to the technical field of plate-type product production, in particular to a method and a system for joint optimization of production scheduling, batch-sample scheduling of plate-type products, wherein the method comprises the steps of constructing a production scheduling model based on a column generation algorithm, decomposing main problems and sub problems, iteratively excavating a negative reduction cost scheme column, and generating a production scheduling result meeting operation coverage and productivity constraint; based on time slices of the scheduling result, constructing and normalizing multi-dimensional feature vectors of orders, adopting weighted Euclidean distance to measure similarity, generating a batch result of compatible batches through multi-constraint clustering and validity screening, adopting a double iterative search mechanism and a variable domain search algorithm based on the batch result, combining a post-processing strategy to generate a batch scheme with minimum unset area, and simultaneously realizing information sharing and bidirectional adjustment among three stages in the optimization process. The invention can solve the bottleneck of sectional solving and difficult integration in the prior art, and improves the execution efficiency and the resource utilization level of the plate-type product manufacturing process.

Inventors

WEI LIJUN
WANG HONGWEI
ZHANG HAO
YAO SHAOWEN
Shu Wenlan
LIU QIANG

Assignees

广东工业大学

Dates

Publication Date: 20260508
Application Date: 20250922

Claims (6)

1. The combined optimization method for the production scheduling, batch-sample arrangement of the plate-type products is characterized by comprising the following steps of: s1, a task-driven multi-constraint scheduling stage, namely constructing a scheduling model based on a column generation algorithm, decomposing a main problem and a sub problem, iteratively excavating a negative reduction cost scheme column, and generating a scheduling result meeting the constraint of job coverage and productivity; s2, an order batch optimization stage under constraint control comprises the steps of constructing order multidimensional feature vectors based on time slices of a batch production result, normalizing, measuring similarity by adopting weighted Euclidean distance, and generating a batch result compatible with the batch production through multi-constraint clustering and validity screening; S3, a heuristic layout optimization stage driven by layout performance, namely generating a layout scheme with minimum unset area by adopting a double iteration search mechanism and a variable domain search algorithm and combining a post-processing strategy based on a batch result; S4, three-stage linkage and closed loop feedback, namely guiding batch production results to generate batch production results by using the batch production results, inputting the batch production results into a layout to generate a layout scheme, and if the layout scheme has abnormal layout, feeding back to a scheduling stage or a batch optimization stage for adjustment, so as to finally generate the layout scheme with minimum unset area meeting the requirements; The scheduling model based on the column generation algorithm in S1 comprises symbol definition, namely a set J of production jobs J to be scheduled, a set S of feasible production schemes S, a scheme selection variable x s , a job coverage variable a s,j , a lower limit UB of the number of machine use and an upper limit LB of the number of machine use; Wherein J e J, S e S, x s ∈｛0,1｝,a s,j e {0,1}; when the feasible scheduling scheme S epsilon S is selected, x s =1, otherwise, 0; when the feasible production schedule s covers the production job j to be scheduled, a s,j =1, otherwise 0; The objective function of the main problem in S1 is: Constraints of the main problem include coverage constraints, variable value constraints and machine use upper and lower bound constraints: the reduced cost calculation formula for the sub-problem described in S1 is: Wherein a j is the implicit cost weight of the production job j to be scheduled, and β is the implicit cost weight of the number of machines, if r s <0, the solution is a negative reduced cost solution column; the objective function of the sub-problem is: in S3, the variable domain search algorithm is used for decoding the raw material plate use sequence, the post-processing strategy is used for further improving the current solution, and the double iteration search mechanism is used for dynamically controlling the iteration number of the variable neighborhood search and the range of the search combination, and the specific steps are as follows: Dividing the plate combination into a plurality of groups according to the sum of available areas of the raw material plates by a double iterative search mechanism, wherein the objective function values of the combinations in each group are the same, searching for a feasible layout scheme for each group, namely, if a feasible scheme is found, cutting out all combinations with target values not less than the group, reducing the search space, and if not found, expanding the combination range and continuing searching; The variable neighborhood search algorithm firstly generates an initial sequence, namely an initial part placement sequence is generated by sorting the parts from large to small, secondly generates a candidate solution, namely a new candidate solution sequence is generated by carrying out random oscillation on the current sequence, and then optimizes local search, namely a local search algorithm LSearch is called to optimize the candidate solution, LSearch adopts three neighborhood structures, and the steps are sequentially executed: A neighborhood 1, randomly exchanging the positions of any two parts; a neighborhood 2, randomly selecting a part, and inserting the part into a random position in the sequence; selecting any local subsequence and executing turning operation; and the searching logic is used for immediately updating the current solution and returning to the neighborhood 1 for searching again if the current neighborhood finds a better solution, switching to the next neighborhood if the current neighborhood is not improved, and stopping and returning to the best solution if all the neighborhood is not improved.
2. The combined optimization method of plate-type products according to claim 1, wherein in S1, the sub-problems are solved by a dynamic programming algorithm based on label setting, the dynamic programming algorithm introduces a dominant rule and upper bound estimation, label retention with lower cumulative dual value is performed under the same residual capacity, and if a part of decomposition cannot obtain negative reduction cost, the method is terminated in advance.
3. The combined production-batching-sample-discharging optimization method for board-like products according to claim 1, wherein in S2, the order multidimensional feature vector includes product geometry, board material, color coding, surface process category, process path label, delivery deadline, priority; the normalization is Min-Max normalization or Z-score normalization.
4. The method for combined optimization of production, batch and layout of board-like products according to claim 1, wherein the similarity measure in S2 uses weighted euclidean distance: Wherein ω k ε [0,1] is the weight of the kth dimension feature, and ; The criteria for multi-constraint clustering and legitimacy screening comprise unified batch inner plate types, the number of product types in the batch is less than or equal to a set value, the span of the batch exchange period is less than or equal to a threshold value, the similarity of process paths in the batch is greater than or equal to a threshold value, and the total order area/number of batches is greater than or equal to a lower limit value.
5. The method for combined optimization of sheet product scheduling, batch and layout according to claim 1, wherein the triggering conditions of the closed loop feedback in S4 include that the layout unset area is greater than or equal to a set threshold, the batch process is incompatible, the sheet material requirement is not matched with stock, the task rhythm is not matched with machine load, and the feedback adjustment mode includes batch splitting/merging, order reassignment, job-machine assignment adjustment and task online time adjustment.
6. A combined production-batch-discharge optimization system for plate products, characterized in that a combined production-batch-discharge optimization method for plate products according to any one of claims 1-5 is adopted, comprising: the scheduling optimization module is used for constructing a scheduling model based on a column generation algorithm, and iteratively excavating a negative reduction cost scheme column through main problem and sub problem decomposition to generate a scheduling result meeting the job coverage and productivity constraint; The batch optimizing module is used for constructing and normalizing the multi-dimensional feature vector of the order based on the time slices of the scheduling result, measuring the similarity by adopting the weighted Euclidean distance, and generating a batch result compatible with the scheduling through multi-constraint clustering and validity screening; The stock layout optimization module is used for generating a stock layout scheme with the minimum unset area by adopting a double iteration search mechanism and a variable domain search algorithm and combining a post-processing strategy based on the batch result; The combined optimization module is used for guiding the batch generation of the batch result by using the batch production result, inputting the batch result into the batch to generate a batch discharging scheme, and feeding back to the batch production optimization module or the batch optimization module for adjustment if the batch discharging scheme has abnormal discharging, so as to finally generate the batch discharging scheme with minimum unset area meeting the requirement.

Description

Combined optimization method and system for scheduling production, batch and sample arrangement of plate-type products Technical Field The invention relates to the technical field of plate-type product production, in particular to a method and a system for joint optimization of production, batch and sample arrangement of plate-type products. Background With the continuous increase of flexible manufacturing and personalized customization demands, the manufacturing process of plate-type products (mainly using plates as main raw materials, adopting plane processing as main forms and being assembled by various plate-type accessories) presents the characteristics of multi-batch, small order and high-frequency switching. The manufacturing process is required to sequentially complete three key stages of task scheduling, order batch and layout, and tight information circulation, resource matching and structure regulation and control association exist among the stages, wherein the scheduling is required to balance the rationality of the structure and the layout efficiency, and the layout is required to realize the maximum utilization of materials and match the batch structure and the task online sequence. However, the prior art generally adopts an optimization strategy of "phase separation and sequential execution", and has the following core drawbacks: (1) The production scheduling and subsequent links are disjointed, namely the production scheduling result is directly used for batch and layout as static input, and the influence of capacity allocation, order online time and delivery cycle on the subsequent links is not considered in the initial period of optimization, so that the production rhythm is unbalanced; (2) The flexibility of the batch strategy is insufficient, namely the batch rule is immobilized, the batch structure cannot be dynamically adjusted according to the product attribute (such as materials and process paths) and production, and the problems of poor compatibility in batches and increased layout difficulty are easy to occur; (3) The layout has strong independence, namely the layout optimization is independent of the production and batch stages, only the material utilization rate is pursued, and the process feasibility and task rhythm matching are neglected, so that the layout scheme cannot adapt to the actual production plan; (4) The global optimization capability is lost, namely the highly-coupled relation of scheduling, batched and stock-removing is difficult to model uniformly in the face of complex heterogeneous orders, the calculation complexity rises sharply along with the expansion of the task scale, the problems of overlong solving time and unstable results occur, and the overall scheduling quality and the real-time response capability are restricted. Disclosure of Invention The invention aims to provide a combined optimization method for scheduling production, batch and sample arrangement of plate-type products, which can break a stage barrier and realize multi-loop cooperation, so that the bottleneck of sectional solution and difficult integration in the prior art is solved, and the execution efficiency and the resource utilization level of the plate-type product manufacturing process are improved. Another object of the present invention is to provide a combined optimizing system for discharging, batching and discharging of plate products, which adopts the combined optimizing method for discharging, batching and discharging of plate products as described above. To achieve the purpose, the invention adopts the following technical scheme: a combined optimization method for production scheduling, batch-sample arrangement of plate-type products comprises the following steps: s1, a task-driven multi-constraint scheduling stage, namely constructing a scheduling model based on a column generation algorithm, decomposing a main problem and a sub problem, iteratively excavating a negative reduction cost scheme column, and generating a scheduling result meeting the constraint of job coverage and productivity; s2, an order batch optimization stage under constraint control comprises the steps of constructing order multidimensional feature vectors based on time slices of a batch production result, normalizing, measuring similarity by adopting weighted Euclidean distance, and generating a batch result compatible with the batch production through multi-constraint clustering and validity screening; S3, a heuristic layout optimization stage driven by layout performance, namely generating a layout scheme with minimum unset area by adopting a double iteration search mechanism and a variable domain search algorithm and combining a post-processing strategy based on a batch result; And S4, three-stage linkage and closed-loop feedback, namely guiding the batch production result to generate a batch production result, inputting the batch production result into a discharging sample to generate a discharging schem