An intelligent dispatching system for cable production lines

Optimizing the shear sequence through dynamic graph construction and state embedding learning modules, the problem of insufficient real-time adaptability and processing accuracy of the cable production line control system is solved, and the material utilization rate and production efficiency are improved, ensuring the flexibility of the production line and efficient operation.

CN118778580BActive Publication Date: 2024-11-19SICHUAN WEIPENG CABLE MFG CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411260250.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-10
Publication Date
2024-11-19
Estimated Expiration
2044-09-10

AI Technical Summary

Technical Problem

The existing cable production line control system has shortcomings in real-time adaptability and processing accuracy, and the production plan cannot be adjusted in time, resulting in waste of materials and overcapacity, low shear errors and sequence efficiency, reducing the overall efficiency and market competitiveness of the production line.

Method used

The dynamic graph building module is used to monitor the cable batch information in real time, the state embedding learning module is used for feature analysis, the shear sequence optimization module optimizes the shear sequence, the production decision optimization module adjusts the production plan, dynamically updates the production process, and improves adaptability and efficiency.

Benefits of technology

Through in-depth analysis of the cable batch characteristics and shear sequence, the shear machine operation plan is optimized, which significantly improves material utilization, reduces production waste, improves production efficiency and quality control, and ensures flexibility and efficient operation of the production line.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118778580B_ABST
    Figure CN118778580B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of production line control technology, specifically to an intelligent scheduling system for a cable production line, the system comprising: a dynamic graph construction module that monitors the entry and exit of cable batches in the production process in real time based on current cable batch information and processed batch information, and updates the dynamic dependency between cable batches and shears. In the present invention, by dynamically updating the production process dependency and deeply analyzing the interactive data, the adaptability and operational efficiency of cable production are effectively improved, the flexibility of production scheduling is significantly enhanced by real-time reflection of production status and batch changes, allowing the production line to quickly adapt to fluctuations in raw materials or demand, in-depth analysis of cable batch characteristics and optimization of shearing sequences significantly improves material utilization, reduces waste in production, and directly improves production efficiency. By optimizing the shearing machine operation plan, the production process becomes more efficient, effectively reducing downtime and production delays, and ensuring high standards of continuous production and quality control.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of production line control, and in particular to an intelligent scheduling system for a cable production line. Background Art

[0002] The field of production line control technology covers various methods and tools that use automation and information technology to optimize production processes. The core of this field is to improve production efficiency, reduce human errors, and achieve cost-effectiveness through advanced software systems and machine automation components. The application of technology includes but is not limited to scheduling algorithms, real-time monitoring, data analysis and predictive maintenance. Production line control systems can be applied to various industrial production environments to ensure the continuity, flexibility and scalability of the production process. With the advancement of Industry 4.0 and smart manufacturing, this field is developing rapidly, introducing more cloud-based services and artificial intelligence technologies to support more complex production tasks and management needs.

[0003] Among them, the intelligent scheduling system for cable production lines refers to a specific type of production line control system used in the cable manufacturing process. The purpose of this system is to optimize the production process through algorithms and automation technology to achieve the best utilization of raw materials, efficient management of the production process and improvement of the quality of the finished product. The intelligent scheduling system can adjust the production plan and machine parameters based on real-time data, thereby reducing downtime and waste and increasing the overall efficiency and output of the production line. This system plays a vital role in maintaining production flexibility and adapting to changes in market demand, especially in the highly competitive cable manufacturing industry.

[0004] Existing technologies have obvious deficiencies in real-time adaptability and processing accuracy. Due to their reliance on static scheduling schemes and restrictive data processing capabilities, conventional production line control systems cannot effectively respond to real-time changes, such as supply chain fluctuations or drastic changes in market demand. This static dependence leads to the inability to adjust production plans in a timely manner, which is common in material waste and overcapacity problems. Traditional systems also show deficiencies in optimizing production batch characteristics and shearing sequences, which often lead to shearing errors and low sequence efficiency, which not only increases production costs but also prolongs the production cycle. These problems reduce the overall efficiency and market competitiveness of the production line, especially in an environment that requires a rapid response to market changes. A production system that lacks flexibility can easily lead to a decline in corporate competitiveness. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an intelligent scheduling system for a cable production line.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A cable production line intelligent scheduling system comprises:

[0007] The dynamic graph construction module monitors the entry and exit of cable batches in the production process in real time based on the current cable batch information and the processed batch information, updates the dynamic dependency between cable batches and shears, adjusts the graph structure to reflect the current production status, and generates a dynamic graph structure;

[0008] The state embedding learning module performs real-time feature analysis on each cable batch node based on the dynamic graph structure, calculates feature embedding values ​​through the interaction data between the cable batch and the shearing machine, obtains low-dimensional state representation of the node, and obtains batch feature embedding;

[0009] The shearing sequence optimization module utilizes the batch feature embedding to compare and analyze the impact of differentiated shearing sequences on production line efficiency, selects the most optimized shearing sequence in real time to improve material utilization and production efficiency, and obtains the optimized shearing sequence;

[0010] The production decision optimization module adjusts the operation plan of the shearing machine and the order of using the cables according to the optimized shearing sequence and the real-time data of the production line, optimizes the production process scheduling, responds to changes in the production process, and establishes an optimized production plan.

[0011] As a further solution of the present invention, the step of obtaining the dynamic graph structure is specifically as follows:

[0012] Collect the current cable batch information, including the joining time and the exit time, using the formula:

[0013]

[0014] Calculate batch timestamps and generate timestamp arrays for cable batches;

[0015] in, Indicates The timestamp of the batch, It is The batch joining time, It is The exit time of each batch, , is the coefficient for adjusting time weighting;

[0016] Using the timestamp array of the cable batches, the dynamic dependencies between all batches are calculated using the formula:

[0017]

[0018] Calculate dependency scores and generate a dependency score matrix;

[0019] in, Representative batch For batches The dependency score of , is the coefficient that adjusts the sensitivity and shape of the function curve, , is an element in the timestamp array of the cable batch;

[0020] Based on the dependency score matrix, the current graph structure is updated using the formula:

[0021]

[0022] Update the graph structure and generate a dynamic graph structure ;

[0023] in, It is a dynamic graph structure. is the old graph structure, Indicates batch With batch Whether there is an edge between them, 1 means yes, 0 means no, is the maximum value in the dependency score matrix.

[0024] As a further solution of the present invention, the step of acquiring the batch feature embedding is specifically as follows:

[0025] Select each cable batch node from the dynamic graph structure, calculate the interaction data between the node and the shearing machine, and use the formula:

[0026]

[0027] Generate node interaction strength vector;

[0028] in, Indicates The interaction strength of batch nodes, It is with The set of nodes directly connected to a node, is the basic weight coefficient of the interaction data, Is a node With Node The connection weights between is the adjustment factor, Indicates the difference in additional attributes between nodes;

[0029] Using the node interaction strength vector, the formula is adopted:

[0030]

[0031] Convert the interaction strength into a feature embedding value and obtain a feature embedding array;

[0032] in, It is The feature embedding values ​​of batch nodes, is the coefficient that adjusts the sensitivity of the nonlinear transformation, is the normalization constant;

[0033] According to the feature embedding array, the formula is:

[0034]

[0035] Perform dimensionality reduction to obtain low-dimensional state representation and generate batch feature embedding;

[0036] in, It is batch feature embeddings for batches, is the number of dimensions after dimensionality reduction, It is The weight coefficient of the dimension, It is The node in The feature embedding value of the dimension.

[0037] As a further solution of the present invention, the step of obtaining the optimized shearing sequence is specifically as follows:

[0038] Using the batch feature embedding, the impact of each shearing sequence on the production line efficiency is calculated using the formula:

[0039]

[0040] Generate a cleavage sequence efficiency score;

[0041] in, Represents the efficiency score of the cutting sequence, represents the weight coefficient, Represents the production parameters of the batch in the shearing sequence, represents the adjustment factor, represents the duration of the cut sequence;

[0042] Comparative analysis of the shear sequence efficiency scores of the differentiated shear sequences was performed using the formula:

[0043]

[0044] Selecting a cutting sequence with a maximum efficiency score from all estimated cutting sequences to obtain a target cutting sequence;

[0045] in, represents the target cutting sequence, represents the selection of the sequence with the maximum efficiency score from all evaluated shear sequences;

[0046] According to the target shearing sequence, the sequence is implemented and the production line parameters are adjusted, using the formula:

[0047]

[0048] Adjust the cutting sequence according to the real-time production line feedback to obtain the optimized cutting sequence;

[0049] in, represents the adjusted shear sequence, represents the adjustment factor, represents the feedback adjustment coefficient, Represents real-time production line status parameters.

[0050] As a further solution of the present invention, the step of obtaining the optimized production plan is specifically as follows:

[0051] Using the optimized shearing sequence and real-time production data, the formula is adopted:

[0052]

[0053] Calculate the adjusted shearing machine parameters and generate the adjusted shearing machine parameter table;

[0054] in, represents the operating coefficient, represents the safety factor, Representative The real-time load of the shearing machine, Represents the set of parameters after shearing machine adjustment;

[0055] Analyze the real-time data and the adjusted shearing machine parameter table, using the formula:

[0056]

[0057] Select the optimal operation plan that reduces the difference in shearing machine parameters, update the shearing machine operation plan, and obtain the optimized operation plan;

[0058] in, Represents optimization of work plan, represents the operation that captures the minimization of the objective function, Indicates The parameter set of the shearing machine after adjustment, represents the probability distribution function;

[0059] Combined with the optimization operation plan and adjustment requirements, the formula is adopted:

[0060]

[0061] Adjust the order of cable use, optimize production process scheduling, and establish an optimized production plan;

[0062] in, represents the optimized production plan, Represents the change in demand over time. represents the demand dynamics after exponential adjustment, Represents the differential of a time interval.

[0063] Compared with the prior art, the advantages and positive effects of the present invention are:

[0064] In the present invention, the adaptability and operational efficiency of cable production are effectively improved by dynamically updating the production process dependencies and deeply analyzing the interactive data. The flexibility of production scheduling is significantly enhanced by reflecting the production status and batch changes in real time, allowing the production line to quickly adapt to fluctuations in raw materials or demand. In-depth analysis of cable batch characteristics and optimization of shearing sequences significantly improves material utilization, while reducing waste in production, directly improving production efficiency. By optimizing the shearing machine operation plan, the production process becomes more efficient, effectively reducing downtime and production delays, and ensuring high standards of continuous production and quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 is a system flow chart of the present invention;

[0066] Figure 2 A flowchart of the steps for obtaining the dynamic graph structure of the present invention;

[0067] Figure 3 A flowchart of the steps for acquiring batch feature embedding of the present invention;

[0068] Figure 4 A flow chart of the steps for obtaining the optimized shearing sequence of the present invention;

[0069] Figure 5 This is a flow chart of the steps for obtaining the optimized production plan of the present invention. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0071] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0072] Embodiment 1

[0073] See also Figure 1 , a cable production line intelligent scheduling system includes:

[0074] The dynamic graph construction module monitors the entry and exit of cable batches in the production process in real time based on the current cable batch information and the processed batch information, updates the dynamic dependency between cable batches and shears, adjusts the graph structure to reflect the current production status, and generates a dynamic graph structure;

[0075] The state embedding learning module performs real-time feature analysis on each cable batch node based on the dynamic graph structure, calculates the feature embedding value through the interaction data between the cable batch and the shearing machine, obtains the low-dimensional state representation of the node, and obtains the batch feature embedding;

[0076] The shearing sequence optimization module uses batch feature embedding to compare and analyze the impact of differentiated shearing sequences on production line efficiency, selects the most optimized shearing sequence in real time to improve material utilization and production efficiency, and obtains the optimized shearing sequence;

[0077] The production decision optimization module adjusts the shearing machine's operation plan and the order of cable usage according to the optimized shearing sequence and real-time data of the production line, optimizes the production process scheduling, responds to changes in the production process, and establishes an optimized production plan.

[0078] The dynamic graph structure specifically includes the joining and exiting of cable batches, dynamic dependencies, and production status feedback. Batch feature embedding includes real-time feature analysis of cable batch nodes, feature embedding values, and low-dimensional state representation. The optimized shearing sequence specifically refers to material utilization and production efficiency. The optimized production plan specifically includes shearing machine operation plan, cable usage sequence, and production process scheduling.

[0079] See also Figure 2 ,The specific steps for obtaining the dynamic graph structure are:

[0080] Collect the current cable batch information, including the joining time and the exit time, using the formula:

[0081]

[0082] Calculate batch timestamps and generate timestamp arrays for cable batches;

[0083] in, Indicates The timestamp of the batch, It is The batch joining time, It is The exit time of each batch, , is the coefficient for adjusting time weighting;

[0084] Using the timestamp array of the cable batches, the dynamic dependencies between all batches are calculated using the formula:

[0085]

[0086] Calculate dependency scores and generate a dependency score matrix;

[0087] in, Representative batch For batches The dependency score of , is the coefficient that adjusts the sensitivity and shape of the function curve, , is an element in the timestamp array of the cable batch;

[0088] Based on the dependency score matrix, update the current graph structure using the formula:

[0089]

[0090] Update the graph structure and generate a dynamic graph structure ;

[0091] in, It is a dynamic graph structure. is the old graph structure, Indicates batch With batch Whether there is an edge between them, 1 means yes, 0 means no, is the maximum value in the dependency score matrix.

[0092] formula:

[0093]

[0094] , : Weighting coefficient for joining and exiting time.

[0095] : No. The time to add the batch.

[0096] : No. The exit time of each batch.

[0097] Setting for batch : Hour, Hour, , ;

[0098] Calculation process:

[0099]

[0100] here The calculation is the cable batch Timestamps in the production process refer to the impact of batch entry and exit times.

[0101] formula:

[0102]

[0103] and : Coefficients that affect the shape and sensitivity of the function.

[0104] and :batch and The timestamp of the

[0105] Set for another batch : Hour, , ;

[0106] Calculation process:

[0107]

[0108] This formula calculates the batch For batches The dependency score reflects the influence of the time difference between the two batches.

[0109] formula:

[0110]

[0111] : Old graph structure.

[0112] :batch right The dependency score.

[0113] The maximum value in the score matrix.

[0114] :batch and Is there a connection between them?

[0115] set up: (no previous connection), (the maximum score calculated previously), (Set all batches to have connections), calculation process:

[0116]

[0117] This formula updates the graph structure, normalizes the dependency scores between each batch, and adds them to the original graph structure.

[0118] See also Figure 3 , the specific steps for obtaining batch feature embedding are:

[0119] Select each cable batch node from the dynamic graph structure and calculate the interaction data between the node and the shearing machine using the formula:

[0120]

[0121] Generate node interaction strength vector;

[0122] in, Indicates The interaction strength of batch nodes, It is with The set of nodes directly connected to a node, is the basic weight coefficient of the interaction data, Is a node With Node The connection weights between is the adjustment factor, Indicates the difference in additional attributes between nodes;

[0123] Use the node interaction strength vector, using the formula:

[0124]

[0125] Convert the interaction strength into a feature embedding value and obtain a feature embedding array;

[0126] in, It is The feature embedding values ​​of batch nodes, is the coefficient that adjusts the sensitivity of the nonlinear transformation, is the normalization constant;

[0127] According to the feature embedding array, the formula is used:

[0128]

[0129] Perform dimensionality reduction to obtain low-dimensional state representation and generate batch feature embedding;

[0130] in, It is batch feature embeddings for batches, is the number of dimensions after dimensionality reduction, It is The weight coefficient of the dimension, It is The node in The feature embedding value of the dimension.

[0131] formula:

[0132]

[0133] : Basic weight coefficient, set to 0.5, indicating the basic weight of interaction data.

[0134] : The connection weight between node i and node j in the dynamic graph structure. Set node i to connect with two adjacent nodes with weights of 1.2 and 0.8 respectively.

[0135] : Adjustment factor, set to 0.3, to enhance the weight influence.

[0136] : Additional attribute differences between nodes, setting the difference values ​​between the same two nodes to 2 and 1.

[0137] Assume that node i has two adjacent nodes j1 and j2.

[0138] for

[0139] for

[0140]

[0141] It means that the total interaction strength of node i is 1.9, which is a result of combining the connection weight and the difference in node attributes.

[0142] formula:

[0143]

[0144] : Nonlinear transformation sensitivity adjustment coefficient, set to 1.0.

[0145] : Normalization constant, set to 1.0.

[0146] : The interaction strength is calculated to be 1.9.

[0147] Example calculation:

[0148]

[0149] Represents the feature embedding value of node i, which is a result adjusted by nonlinear transformation and reflects the characteristics of the node in the dynamic graph.

[0150] formula:

[0151]

[0152] : The weight coefficient of the kth dimension, set .

[0153] : The feature embedding value of the kth dimension, set with Same as 0.870.

[0154] Example calculation:

[0155]

[0156] The low-dimensional state representation of node i is the result of simplifying the feature embedding value to a low-dimensional space through the weight coefficient.

[0157] See also Figure 4 , the specific steps for obtaining the optimized shear sequence are:

[0158] Using batch feature embedding, the impact of each shearing sequence on production line efficiency is calculated using the formula:

[0159]

[0160] Generate a cleavage sequence efficiency score;

[0161] in, Represents the efficiency score of the cutting sequence, represents the weight coefficient, Represents the production parameters of the batch in the shearing sequence, represents the adjustment factor, represents the duration of the cut sequence;

[0162] Comparative analysis of the shearing sequence efficiency scores of differentially sheared sequences was performed using the formula:

[0163]

[0164] Selecting a cutting sequence with a maximum efficiency score from all estimated cutting sequences to obtain a target cutting sequence;

[0165] in, represents the target cutting sequence, represents the selection of the sequence with the maximum efficiency score from all evaluated shear sequences;

[0166] According to the target shearing sequence, implement the sequence and adjust the production line parameters, using the formula:

[0167]

[0168] Adjust the cutting sequence according to the real-time production line feedback to obtain the optimized cutting sequence;

[0169] in, represents the adjusted shear sequence, represents the adjustment factor, represents the feedback adjustment coefficient, Represents real-time production line status parameters.

[0170] formula:

[0171]

[0172] is the weight coefficient, reflecting the importance of differentiated batches.

[0173] is the square of the batch feature embedding, which is used to emphasize the impact of feature embedding.

[0174] is the production parameter of the batch in the target shear sequence.

[0175] is the adjustment factor used to adjust the duration of the cut sequence impact.

[0176] There are three batches with weights They are .

[0177] Feature Embedding They are .

[0178] Production parameters They are .

[0179] Adjustment factor is 0.05.

[0180] Duration They are .

[0181] calculate :

[0182]

[0183] The value (28.865) represents the overall efficiency score of the sheared sequence after considering the feature embedding and production parameters.

[0184] formula:

[0185]

[0186] Several different sequence schemes were set up and compared, and the results were The highest is currently calculated at 28.865.

[0187] formula:

[0188]

[0189] Adjustment factor is 1.

[0190] Feedback adjustment factor is 0.1.

[0191] Real-time production line status parameters is 400.

[0192] calculate:

[0193]

[0194] This value (30.865) represents the efficiency score of the trimming sequence adjusted after referring to the real-time feedback.

[0195] See also Figure 5 , the specific steps for obtaining the optimized production plan are:

[0196] Using the optimized shear sequence and real-time production data, the formula:

[0197]

[0198] Calculate the adjusted shearing machine parameters and generate the adjusted shearing machine parameter table;

[0199] in, represents the operating coefficient, represents the safety factor, Representative The real-time load of the shearing machine, Represents the set of parameters after shearing machine adjustment;

[0200] Analyze the real-time data and the adjusted shearing machine parameter table, using the formula:

[0201]

[0202] Select the optimal operation plan that reduces the difference in shearing machine parameters, update the shearing machine operation plan, and obtain the optimized operation plan;

[0203] in, Represents optimization of work plan, represents the operation that captures the minimization of the objective function, Indicates The parameter set of the shearing machine after adjustment, represents the probability distribution function;

[0204] Combined with the optimization of operation plan and adjustment requirements, the formula is adopted:

[0205]

[0206] Adjust the order of cable use, optimize production process scheduling, and establish an optimized production plan;

[0207] in, represents the optimized production plan, Represents the change in demand over time. represents the demand dynamics after exponential adjustment, Represents the differential of a time interval.

[0208] Formula analysis

[0209]

[0210] : Operation coefficient, representing the influence weight of multiple shears.

[0211] : The optimized sequence value of the i-th shear.

[0212] : Safety factor, affecting the adjustment sensitivity to real-time load.

[0213] : Real-time load data of the i-th shearing machine.

[0214] There are two shearing machines, and the relevant parameters are set as follows:

[0215] For all shears, are the sequence values ​​of the two shears respectively

[0216] , Real-time load data;

[0217] Calculate the parameter contribution of each shear:

[0218] For the first shear: ;

[0219] For the second shear: ;

[0220] therefore, ;

[0221] Formula analysis:

[0222]

[0223] There are two potential job schedules and , it is necessary to determine which plan minimizes the variance.

[0224] For P1:

[0225] The first shearing machine: ;

[0226] Second shear: ;

[0227] Total difference: ;

[0228] For P2:

[0229] The first shearing machine: ;

[0230] Second shear: ;

[0231] Total difference: ;

[0232] therefore, is P2, because there is a smaller difference (0.62).

[0233] Formula analysis:

[0234]

[0235] Set within a time period The value is a constant 0.1, and set If the value in lasts for one unit of time, the order of cable use is adjusted as follows:

[0236]

[0237] Therefore, the adjusted usage order is approximately [2.21,5.525].

[0238] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent scheduling system for a cable production line, characterized in that: The system includes a dynamic graph construction module, a state embedding learning module, a cutting sequence optimization module, and a production decision optimization module: The dynamic graph construction module monitors the entry and exit of cable batches in the production process in real time based on the current cable batch information and the processed batch information, updates the dynamic dependency between the cable batches and the shearing machine, adjusts the graph structure to reflect the current production status, and generates a dynamic graph structure; The steps for obtaining the dynamic graph structure are specifically as follows: Collect the current cable batch information, including the joining time and the exit time, using the formula: T(i)=λ 1o a i +λ 2o d i Calculate batch timestamps and generate timestamp arrays for cable batches; Among them, T(i) represents the timestamp of the i-th batch, a i is the time of adding the i-th batch, δ i is the exit time of the i-th batch, λ 1o , 2o is the coefficient for adjusting time weighting; Using the timestamp array of the cable batches, the dynamic dependencies between all batches are calculated using the formula: Calculate dependency scores and generate a dependency score matrix; Among them, D ij represents the dependency score of batch i on batch j, γ o , β o are coefficients for adjusting the sensitivity and shape of the function curve, T(i) and T(j) are elements in the timestamp array of the cable batch; Based on the dependency score matrix, the current graph structure is updated using the formula: Update the graph structure and generate a dynamic graph structure G new ; Among them, G new is a dynamic graph structure, G old is the old graph structure, E ij Indicates whether there is an edge between batch i and batch j, 1 means yes, 0 means no, D max is the maximum value in the dependency score matrix; The state embedding learning module performs real-time feature analysis on each cable batch node based on a dynamic graph structure, calculates feature embedding values ​​through interaction data between the cable batch and the shearing machine, obtains low-dimensional state representation of the node, and obtains batch feature embedding; The shearing sequence optimization module uses batch feature embedding to compare and analyze the impact of differentiated shearing sequences on production line efficiency, selects the most optimized shearing sequence in real time to improve material utilization and production efficiency, and obtains the optimized shearing sequence; The production decision optimization module adjusts the shearing machine's operation plan and the order of cable usage according to the optimized shearing sequence and real-time data of the production line, optimizes the production process scheduling, responds to changes in the production process, and establishes an optimized production plan.

2. The intelligent dispatching system for cable production line according to claim 1 is characterized in that: The steps for obtaining the batch feature embedding are specifically as follows: Select each cable batch node from the dynamic graph structure, calculate the interaction data between the node and the shearing machine, and use the formula: V i =∑ j∈N(i) a u ·G new (i,j)+μ u ·∑ j∈N(i) d ij Generate node interaction strength vector; Among them, V i represents the interaction strength of the i-th batch of nodes, N(i) is the set of nodes directly connected to the i-th node, α u is the basic weight coefficient of the interaction data, G new (i,j) is the connection weight between node i and node j, μ u is the adjustment factor, δ ij Indicates the difference in additional attributes between nodes; Using the node interaction strength vector, the formula is adopted: Convert the interaction strength into a feature embedding value and obtain a feature embedding array; Among them, E i is the feature embedding value of the i-th batch node, β u is the coefficient for adjusting the sensitivity of nonlinear transformation, κ u is the normalization constant; According to the feature embedding array, the formula is: Perform dimensionality reduction to obtain low-dimensional state representation and generate batch feature embedding; Among them, F i is the batch feature embedding of the i-th batch, K is the number of dimensions after dimensionality reduction, is the weight coefficient of the kth dimension, is the feature embedding value of the i-th node in the k-th dimension.

3. The intelligent dispatching system for cable production line according to claim 2 is characterized in that: The steps for obtaining the optimized shearing sequence are specifically as follows: Using the batch feature embedding, the impact of each shearing sequence on the production line efficiency is calculated using the formula: Generate a cleavage sequence efficiency score; Among them, S eff represents the efficiency score of the shearing sequence, w y Represents the weight coefficient, P y represents the production parameters of the batch in the shearing sequence, ξ y represents the adjustment factor, D y represents the duration of the cut sequence; Comparative analysis of the shear sequence efficiency scores of the differentiated shear sequences was performed using the formula: Opt seq =argmax seq (S eff ) Selecting a cutting sequence with a maximum efficiency score from all estimated cutting sequences to obtain a target cutting sequence; Among them, Opt seq represents the target cutting sequence, argmax seq (S eff ) represents the selection of the sequence with the maximum efficiency score from all evaluated shear sequences; According to the target shearing sequence, the sequence is implemented and the production line parameters are adjusted, using the formula: Adjust the cutting sequence according to the real-time production line feedback to obtain the optimized cutting sequence; Among them, Adjust seq represents the adjusted shear sequence, λ y represents the adjustment coefficient, θ y Represents the feedback adjustment coefficient, R y Represents real-time production line status parameters.

4. The intelligent dispatching system for cable production line according to claim 3 is characterized in that: The steps for obtaining the optimized production plan are specifically as follows: Using the optimized shearing sequence and real-time production data, the formula is adopted: R params =∑ i c t ·Adjust seq (i) 2 +σ t ·log(L t (i)+1) Calculate the adjusted shearing machine parameters and generate the adjusted shearing machine parameter table; Among them, c t represents the operating coefficient, σ t Represents the safety factor, L t (i) represents the real-time load of the i-th shearing machine, R params Represents the set of parameters after shearing machine adjustment; Analyze the real-time data and the adjusted shearing machine parameter table, using the formula: Select the optimal operation plan that reduces the difference in shearing machine parameters, update the shearing machine operation plan, and obtain the optimized operation plan; Among them, P opt Represents the optimization operation plan, argmin P represents the operation that captures the objective function minimization, R params (i) represents the set of parameters after adjustment of the i-th shearing machine, and P(i) represents the probability distribution function; Combined with the optimization operation plan and adjustment requirements, the formula is adopted: Adjust the order of cable use, optimize production process scheduling, and establish an optimized production plan; Among them, Cable order represents the optimized production plan, D t Represents the change in demand over time. represents the demand dynamics after exponential adjustment, and dt represents the differential of the time interval.

Citation Information

Patent Citations

  • Intelligent production scheduling management system of cable production and processing workshop

    CN117196212A

  • Cable processing quality evaluation method based on intelligent manufacturing production process

    CN117350585A

  • Slitting method of flat cable, and its device

    JP2008171758A

  • Causal relational artificial intelligence and risk framework for manufacturing applications

    US20220342371A1

  • Temporal co-graph machine learning networks

    US20230214720A1