Steel cutting and production scheduling method and device based on dynamic risk assessment

Abstract

The invention provides a steel cutting and production scheduling method and device based on dynamic risk assessment, and relates to the field of steel processing pre-production management. Order, equipment and inventory data are collected in real time, and a multi-dimensional numerical matrix is constructed to achieve data structured expression; adopting a Monte Carlo method to generate a candidate production scheduling sequence, combining leftover material loss and delay risk dynamic evaluation, and fusing simulated annealing and an adaptive threshold mechanism to carry out iterative optimization; and through probability similarity screening and regression fitting, an optimal production scheduling scheme is determined in an error allowable range. The method effectively balances the production efficiency and the resource utilization rate, reduces the cutting loss and the delivery risk, and is suitable for the intelligent production scheduling decision of a complex manufacturing scene.

Description

Technical Field

[0001] The present invention relates to the fields of industrial software and artificial intelligence, and in particular to a steel cutting production scheduling method and device based on dynamic risk assessment. Background Art

[0002] In the steel shearing industry, the quality of production scheduling has a decisive impact on material utilization, production efficiency, and order delivery capabilities. However, existing production scheduling technology systems have significant limitations and are unable to adapt to the diverse, small-batch, and highly flexible production needs of modern steel processing companies.

[0003] Traditional production scheduling systems suffer from significant data processing flaws, failing to effectively integrate and matrix key data such as order parameter sets, equipment state vectors, inventory balance distribution, and material substitution rule tables. Information such as order geometry (length, width, thickness, quantity), priority weights, equipment's real-time throughput, task load, maintenance status, wear index, and inventory coil remaining size, material type, storage location, and substitution rules is often managed in a decentralized manner. This results in low data utilization and a failure to provide a comprehensive and accurate basis for scheduling decisions.

[0004] At the algorithmic level, mainstream scheduling methods lack efficient search capabilities within a constrained space, like Monte Carlo methods, when generating candidate production sequences. Furthermore, their comprehensive assessment of scrap loss and delay risks is inaccurate. Furthermore, during sequence optimization, it's difficult to effectively combine simulated annealing algorithms and adaptive loss threshold mechanisms for crossover and mutation operations, leading to the tendency to fall into local optima and inability to achieve global optimization.

[0005] In terms of dynamic responsiveness, existing systems struggle to adjust production schedules in real time based on changes in production data. When order priorities, equipment status, or inventory levels change, the multidimensional numerical matrix cannot be updated promptly to regenerate the optimal production sequence. This causes the production schedule to be disconnected from actual production scenarios, impacting production efficiency and the timeliness of order delivery.

[0006] In addition, the existing technology has shortcomings in the screening and fitting of production scheduling sequences. It lacks an effective mechanism to eliminate redundant sequences through probability similarity screening and use regression fitting to determine the optimal solution within the allowable error range, so the accuracy and reliability of production scheduling results need to be improved.

[0007] The combined effects of these technical bottlenecks have led to significant pain points in the steel processing industry, including low average material utilization, insufficient on-time order delivery rates, and excessive equipment idle time. Overcoming these technical limitations and building an intelligent production scheduling system capable of matrix processing multi-source data and dynamic global optimization has become an urgent need to promote industry transformation and upgrading. Summary of the Invention

[0008] The main purpose of the present invention is to overcome the defects of the existing steel cutting scheduling technology in the prior art, such as insufficient data integration, single optimization strategy, weak global optimization ability and poor dynamic convergence, and propose a steel cutting scheduling method and device based on dynamic risk assessment. By converting the scheduling data into a multidimensional numerical matrix, integrating the Monte Carlo method to generate a candidate sequence, combining simulated annealing with the adaptive loss threshold mechanism for iterative optimization, and through similarity screening and regression fitting, the dynamic convergence and optimal determination of the scheduling plan are achieved.

[0009] The present invention adopts the following technical solutions: A steel cutting production scheduling method based on dynamic risk assessment includes the following steps: S1 acquires production scheduling data in real time, the production scheduling data including an order parameter set, an equipment state vector, and an inventory balance; performs numerical processing on the production scheduling data and converts it into a multidimensional numerical matrix, the multidimensional numerical matrix including an order feature matrix, an equipment capacity matrix, and a surplus inventory matrix; S2 uses the Monte Carlo method to calculate the order feature matrix, the equipment capacity matrix, and the surplus inventory matrix within the constraint space to generate a candidate production schedule that satisfies the order priority and equipment constraints. For each order in the candidate production schedule, it dynamically calculates scrap loss and delay risk to comprehensively obtain a comprehensive loss assessment value for each order. S3 combines the simulated annealing algorithm and the adaptive loss threshold mechanism to perform crossover and mutation operations on the candidate production scheduling sequences; eliminates redundant sequences through similarity screening, and uses regression fitting to determine the potential optimal production scheduling plan within the allowable error range; and iteratively executes the sequence crossover, mutation, screening and fitting process until the optimal production scheduling plan that meets the preset conditions is obtained.

[0010] The production scheduling data includes the order parameter set , device state vector , inventory balance distribution and material replacement rules table ;in, is the total number of orders currently to be scheduled. For the An order data object; is the total number of available production equipment, For the Status data of each device; is the total number of coil types in stock, For the Inventory data of coil-like materials; Based on the order parameter set, we extract the geometric parameters of the order, including the length, width, thickness, and quantity of the steel required for the order. We also mark the order priority weights and dynamically adjust them based on the delivery urgency and customer level. The order feature matrix we construct is as follows: ; in, is the order number, The length of steel required for the order, The width of steel required for the order, The quantity of steel required for the order, is the priority weight; Based on the device state vector, the real-time throughput, current task load, and maintenance status of the device are obtained. The wear index of the device is calculated and the remaining service life is predicted by combining historical data. The device capability matrix is ​​constructed as follows: ; in, is the device number, is the order length, is the maximum processing length, is the current task queue length; Based on the inventory balance distribution, the remaining size, material type, and storage location of the coils in each warehouse are recorded. The recommended usage scenarios and replacement rules of the remaining materials are marked. The remaining material inventory matrix is ​​constructed as follows: ; in, The coil number, is the ascending length, Number of replaceable material.

[0011] Step S2 specifically includes the following: S21 sets constraint space parameters, including order priority sorting rules, equipment operation thresholds, and production resource constraints; performs random sampling on the order feature matrix, equipment capability matrix, and surplus inventory matrix based on the Monte Carlo method to generate an initial production scheduling sequence within the constraint space; S22 performs compliance verification on the initial production scheduling sequence based on the order priority sorting rules and equipment operation thresholds, eliminates sequences that do not meet the constraints, and retains candidate production scheduling sequences that meet the requirements; S23 calculates the scrap loss value for each order in the candidate scheduling sequence by combining the material demand parameters in the surplus inventory matrix and the order feature matrix; and calculates the delay risk coefficient for each order based on the equipment load parameters in the equipment capacity matrix and the delivery date parameters in the order feature matrix. S24 combines the scrap loss value and the delay risk coefficient according to the preset weight to obtain the comprehensive loss assessment value of each order.

[0012] In step S21, the Monte Carlo search space is defined as follows: ; in, represents the Monte Carlo search space, i.e., the set of candidate production schedules. Indicates the Candidate scheduling sequences, represents the full permutation of the order set, Indicates the index of the currently generated sequence, Indicates the maximum number of generated sequences.

[0013] In step S24, the comprehensive loss assessment value is calculated as follows: From the rest stock matrix Filter the coils that meet the conditions: ; in, For a collection of coils that meet the conditions, For the A roll of material, is the remaining length of the coil, The replaceable material number of the coil, The length of steel required for the order, The material type of the order, It is the material replacement rule function; Select the optimal coil material through the value function: ; in, is the optimal coil material selection function, In order to find the coil object with the minimum comprehensive cost in the candidate list, For size matching, is the transportation distance from the warehouse to the processing equipment, To calculate transportation costs; For successfully matched orders , calculate the scrap loss value: ; in, The length of the selected coil, It is The order length of the sub-cut, The number of orders cut from the coil; Processing start time prediction: ; in, is the current system time, is the device queue load rate, For equipment Real-time throughput, For equipment The length of the task queue, is the unit time basis; Delay risk factor calculation: ;in, The processing time for the order is The delivery deadline for the order; Calculate the The comprehensive loss evaluation value of the iteration : ; in, For material consumption, is the time consumption item, is the delay penalty coefficient; ; in, Indicates the number of urgent orders, The total number of orders.

[0014] Step S3 specifically includes the following: S31 initializes simulated annealing parameters, including initial temperature, cooling coefficient and termination temperature, and sets an initial reference value of the adaptive loss threshold; S32 selects a candidate production scheduling sequence with a better comprehensive loss assessment value as the parent generation, fixes the position of the high-priority order, and performs a region crossover operation on the remaining orders to generate a child sequence; randomly swaps the positions of the child sequences obtained by the region crossover to generate a new candidate production scheduling sequence; S33 combines an adaptive loss threshold mechanism with a simulated annealing algorithm to evaluate and accept new candidate production schedules. It then screens candidate production schedules through probability similarity calculations, retaining those that meet the requirements. Based on the steel length error range, it performs scatter point fitting regression calculations on the screened sequences to determine the potential optimal production schedule. S34 reduces the simulated annealing temperature and repeats the above S32 to S33 until the termination condition is met, and finally determines the optimal production scheduling plan; if the requirements are not met, returns to S31 to readjust the parameters and iterate.

[0015] Step S32 includes: Take two candidate production scheduling sequences with better comprehensive loss assessment values ​​as parent sequences, fix the positions of orders with priority ≥ 4, and perform region intersection and random position exchange on the remaining orders to generate subsequences as new candidate production scheduling sequences: Region crossover operation: Randomly select the crossover interval [start, end] from the two parent sequences, where start is the starting position order of the parent sequence and end is the ending position order of the parent sequence. The child sequence directly inherits the interval fragment of one of the parent sequences, and the remaining positions are filled with orders with priority ≤ 3 according to the order of the other parent sequence; Random position swap: Perform n swaps on each subsequence, and only swap the positions of orders with priority ≤ 3, n = 1, 2, 3, ... 5; Combined with the simulated annealing mechanism, the suboptimal solution is accepted with probability. Specifically, the annealing temperature update formula is: ; in, For the The annealing temperature at the iteration, is the initial temperature of the scheduling process, is the decay rate, is the number of iterations; The probability of accepting a suboptimal solution uses the Metropolis criterion: ; in, is the probability of accepting a suboptimal solution, is the difference between the loss evaluation values ​​of the new solution and the current solution, For the The annealing temperature at the iteration.

[0016] Set the adaptive loss threshold and gradually tighten the standard during iteration, specifically: Initial loss threshold ,in is the historical average loss rate; Loss threshold varies with the number of iterations Adaptive Tightening: , ; Set dynamic acceptance probability, allowing the algorithm to accept inferior solutions with a certain probability ( > ), avoid falling into the local optimum, gradually reduce the probability of accepting inferior solutions as the iteration proceeds, and finally stabilize at the optimal solution: ; in, is the comprehensive loss value of the current iteration, is the historical optimal loss value, temperature parameter With the number of iterations attenuation: , is the initial temperature.

[0017] In step S33, candidate production scheduling sequences are screened by probability similarity, specifically including: Collect order information in the candidate scheduling sequence, including material width, length, and quantity; Generate all possible production scheduling combinations based on material length, quantity, and width; Normalize each candidate production scheduling combination and calculate the Euclidean norm with actual demand: If two candidate production scheduling combinations Combined with candidate production schedule The production scheduling is considered successful if the following conditions are met: ; in, Candidate scheduling combination The nth cutting group in Candidate scheduling combination The nth cutting group in For order length information, symbol In order to normalize the three-dimensional feature vector of the candidate production scheduling combination, the symbol is the Euclidean norm calculation, The difference vector between the normalized order demand feature vector and the normalized feature vector of the candidate scheduling combination plan C2; Set similarity threshold : ; in, is the dynamic adjustment factor, ; If the similarity threshold Greater than probability similarity , then the two candidate production scheduling combinations are eliminated, otherwise the two candidate production scheduling combinations are retained; Based on the error range of steel length, scatter fitting regression calculation is performed to obtain the optimal production scheduling plan: Perform scatter point fitting on the set of candidate production schedules after screening, taking into account the 10% length error range of the steel after cutting, to obtain a scatter point fitting set;

[0018] in, =Theoretical length of steel ordered , = Actual cutting length deviation rate ; Length error range constraints: ; Use the least squares method to determine the optimal production scheduling plan, minimizing the error between the actual production scheduling results and the ideal production scheduling plan. Specifically, this includes: a) Construct the input matrix D, which contains the theoretical length of steel for the order Deviation rate from actual cutting length ; b) solving the weighted least squares equation to obtain the deviation prediction model; c) Select the candidate production scheduling sequence that minimizes the absolute value of the forecast deviation as the optimal production scheduling plan.

[0019] The calculation formula of the length deviation rate is as follows: Length deviation rate ; in, The length of steel required for the order, The actual cutting length.

[0020] A steel shearing and production scheduling device based on dynamic convergence, comprising: A data acquisition and processing module acquires production scheduling data in real time, the production scheduling data including order parameter sets, equipment state vectors, and inventory balances; performs numerical processing on the production scheduling data and converts it into a multidimensional numerical matrix, the multidimensional numerical matrix including an order feature matrix, an equipment capacity matrix, and a surplus inventory matrix; The dynamic risk assessment module, based on the Monte Carlo method, operates on the order feature matrix, the equipment capacity matrix, and the surplus inventory matrix within the constraint space to generate a candidate production schedule that satisfies order priorities and equipment constraints. For each order in the candidate production schedule, it dynamically calculates scrap loss and delay risk to comprehensively determine a comprehensive loss assessment value for each order. The production scheduling plan generation module combines the simulated annealing algorithm and the adaptive loss threshold mechanism to perform crossover and mutation operations on the candidate production scheduling sequences; eliminates redundant sequences through similarity screening, and uses regression fitting to determine the potential optimal production scheduling plan within the allowable error range; and iteratively executes the sequence crossover, mutation, screening and fitting process until the optimal production scheduling plan that meets the preset conditions is obtained.

[0021] From the above description of the present invention, it can be seen that compared with the prior art, the present invention has the following beneficial effects: (1) By constructing the order feature matrix, equipment capacity matrix and surplus inventory matrix, data structured integration is achieved. The Monte Carlo method is combined to generate candidate production sequences and dynamically calculate the comprehensive loss assessment value, which significantly improves material utilization. The adaptive loss threshold is tightened with iteration, effectively avoiding the premature convergence problem caused by a fixed threshold.

[0022] (2) Optimize equipment allocation based on parameters such as real-time load and wear index in the equipment capacity matrix to achieve load balancing to avoid overloading of a single device. Dynamically adjust the processing plan based on the equipment capacity to extend the equipment life.

[0023] (3) The mutation strategy of regional crossover and random position exchange is adopted, combined with the probability acceptance mechanism of the simulated annealing algorithm, to effectively avoid local optimal solutions and improve the success rate of global optimization.

[0024] (4) Redundant sequences are eliminated through probabilistic similarity screening, and the optimal solution is determined within the allowable error range by combining regression fitting, thereby improving the accuracy and feasibility of the production scheduling plan; the cyclic iterative mechanism ensures continuous optimization of the plan and further reduces the risk of loss and delay. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is the main flow chart of the method of the present invention; Figure 2 It is a schematic diagram of the basic framework of the present invention; Figure 3 A schematic diagram of a data acquisition feature matrix according to an embodiment of the present invention; Figure 4 Schematic diagram of the annealing mechanism and sequence cross sorting principle of an embodiment of the present invention.

[0026] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. DETAILED DESCRIPTION

[0027] The present invention is further described below through specific embodiments.

[0028] The present invention is an intelligent steel production scheduling method that integrates multi-source data matrix processing, Monte Carlo global optimization and dynamic feedback mechanism. It is mainly used in the intelligent production management system of steel processing enterprises to achieve coordinated optimization of material utilization and order delivery capacity.

[0029] See also Figures 1 to 4 The present invention proposes a steel cutting production scheduling method based on dynamic risk assessment, comprising: S1 acquires production scheduling data in real time, the production scheduling data including order parameter set, equipment state vector and inventory balance; performs numerical processing based on the production scheduling data and converts it into a multidimensional numerical matrix, the multidimensional numerical matrix including order feature matrix, equipment capacity matrix and surplus inventory matrix.

[0030] In this step, the basic structure of data acquisition and matrix construction can be found in Figure 3, through the data interface, the order parameter set (including order number, length, width, quantity, priority weight), equipment status vector (including equipment number, maximum processing length, current task queue length, wear index) and remaining material inventory distribution (including coil number, remaining length, material type, replaceable material number) and material replacement rule table are obtained in real time.

[0031] In the data layer, a multi-source data matrix processing module is used to convert the production scheduling data into a multi-dimensional numerical matrix. Specifically, the order parameter set obtained through the data interface , device state vector , inventory balance distribution and material replacement rules table ;in, is the total number of orders currently to be scheduled. For the An order data object; is the total number of available production equipment, For the Status data of each device; is the total number of coil types in stock, For the Inventory data of coil-like materials.

[0032] Based on the order parameter set, we extract the geometric parameters of the order, including the length, width, thickness, and quantity of the steel required for the order. We also mark the order priority weights and dynamically adjust them based on the delivery urgency and customer level. The order feature matrix we construct is as follows: ; in, is the order number, The length of steel required for the order, The width of steel required for the order, The quantity of steel required for the order, is the priority weight (level 1-5); Based on the device state vector, the real-time throughput, current task load, and maintenance status of the device are obtained. The wear index of the device is calculated and the remaining service life is predicted by combining historical data. The device capability matrix is ​​constructed as follows: ; in, is the device number, is the order length, is the maximum processing length, is the current task queue length; Based on the inventory balance distribution, the remaining size, material type, and storage location of the coils in each warehouse are recorded. The recommended usage scenarios and replacement rules of the remaining materials are marked. The remaining material inventory matrix is ​​constructed as follows: ; in, The coil number, is the ascending length, Number of replaceable material.

[0033] S2 uses the Monte Carlo method to calculate the order feature matrix, the equipment capacity matrix, and the surplus inventory matrix within the constraint space to generate a candidate production schedule that meets the order priority and equipment constraints. For each order in the candidate production schedule, it dynamically calculates scrap loss and delay risk to comprehensively obtain a comprehensive loss assessment value for each order. Specifically, it includes the following: S21 sets constraint space parameters, which include order priority sorting rules, equipment operation thresholds, and production resource constraints; based on the Monte Carlo method, performs random sampling operations on the order feature matrix, equipment capability matrix, and surplus inventory matrix to generate an initial production scheduling sequence within the constraint space.

[0034] In this step, the feature matrix is ​​received 、 、 , generate candidate scheduling sequences , give priority to satisfying the pre-conditions of high-priority orders; define the Monte Carlo search space: ; in, represents the Monte Carlo search space (the set of all possible production schedules), Indicates the Candidate scheduling sequences, represents the full permutation of the order set, Indicates the index of the currently generated sequence, Indicates the maximum number of generated sequences.

[0035] S22 performs compliance verification on the initial production scheduling sequence based on the order priority sorting rules and equipment operation thresholds, eliminates sequences that do not meet the constraints, and retains candidate production scheduling sequences that meet the requirements; S23 calculates the scrap loss value for each order in the candidate scheduling sequence by combining the material demand parameters in the surplus inventory matrix and the order feature matrix; and calculates the delay risk coefficient for each order based on the equipment load parameters in the equipment capacity matrix and the delivery date parameters in the order feature matrix. S24 combines the scrap loss value and the delay risk coefficient according to the preset weight to obtain the comprehensive loss assessment value of each order. The comprehensive loss assessment value is calculated as follows: For each order in the candidate scheduling sequence, perform the following actions: From the rest stock matrix , filter the coils that meet the conditions:

[0036] in, For a collection of coils that meet the conditions, For the A roll of material, is the remaining length of the coil, The replaceable material number of the coil, The length of steel required for the order, The material type of the order, It is the material replacement rule function; Select the optimal coil material through the value function: ; in, is the optimal coil material selection function, In order to find the coil object with the minimum comprehensive cost in the candidate list, For size matching, is the transportation distance from the warehouse to the processing equipment, To calculate transportation costs; For successfully matched orders , calculate the scrap loss value: ; in, The length of the selected coil, It is The order length of the sub-cut, The number of orders cut from the coil; Processing start time prediction: ; in, is the current system time, is the device queue load rate, For equipment Real-time throughput, For equipment The length of the task queue, is the unit time basis; Delay risk factor calculation: ; in, The processing time for the order is The delivery deadline for the order; Calculate the The comprehensive loss evaluation value of the iteration : ; in, For material consumption, is the time consumption item, is the delay penalty coefficient; ; in, Indicates the number of urgent orders, The total number of orders.

[0037] S3 combines the simulated annealing algorithm and the adaptive loss threshold mechanism to perform crossover and mutation operations on the candidate production schedules; eliminates redundant sequences through similarity screening, and uses regression fitting to determine the potential optimal production schedule within the error tolerance range; it iteratively executes the sequence crossover, mutation, screening, and fitting process until the optimal production schedule that meets the preset conditions is obtained. Specifically, it includes the following: S31 initializes simulated annealing parameters, including initial temperature, cooling coefficient and termination temperature, and sets an initial reference value of the adaptive loss threshold.

[0038] S32 selects the candidate production scheduling sequence with the better comprehensive loss assessment value as the parent generation, fixes the position of the high-priority order, and performs a region crossover operation on the remaining orders to generate a child sequence; the child sequence obtained by the region crossover is randomly swapped to generate a new candidate production scheduling sequence. This step specifically includes the following: Take two candidate production scheduling sequences with better comprehensive loss assessment values ​​as parent sequences, fix the positions of orders with priority ≥ 4, and perform region intersection and random position exchange on the remaining orders to generate subsequences as new candidate production scheduling sequences: Region crossover operation: Randomly select the crossover interval [start, end] from the two parent sequences, where start is the starting position order of the parent sequence and end is the ending position order of the parent sequence. The child sequence directly inherits the interval fragment of one of the parent sequences, and the remaining positions are filled with orders with priority ≤ 3 according to the order of the other parent sequence; Random position swap: Perform n swaps on each subsequence, and only swap the positions of orders with priority ≤ 3, n = 1, 2, 3, ... 5; Combined with the simulated annealing mechanism to accept suboptimal solutions with probability, specifically: Annealing temperature update formula: ; in, For the The annealing temperature at the iteration, is the initial temperature of the scheduling process, is the decay rate, is the number of iterations; The probability of accepting a suboptimal solution uses the Metropolis criterion: ; in, is the probability of accepting a suboptimal solution, is the difference between the loss evaluation values ​​of the new solution and the current solution, For the The annealing temperature at the iteration.

[0039] Set the adaptive loss threshold and gradually tighten the standard during iteration, specifically: Initial loss threshold ,in is the historical average loss rate; Loss threshold varies with the number of iterations Adaptive Tightening: , ; Set dynamic acceptance probability, allowing the algorithm to accept inferior solutions with a certain probability ( > ), avoid falling into the local optimum, gradually reduce the probability of accepting inferior solutions as the iteration proceeds, and finally stabilize at the optimal solution: ; in, is the comprehensive loss value of the current iteration, is the historical optimal loss value, temperature parameter With the number of iterations attenuation: , is the initial temperature.

[0040] S33 combines the adaptive loss threshold mechanism with the simulated annealing algorithm to evaluate and accept new candidate production scheduling sequences. It screens the candidate production scheduling sequences through probability similarity calculation and retains the sequences that meet the requirements. Based on the error range of steel length, it performs scatter fitting regression calculation on the screened sequences to determine the potential optimal production scheduling plan.

[0041] In this step, candidate production schedules are screened based on probability similarity, specifically including: Collect order information in the candidate production scheduling sequence, including material width, length, and quantity; generate all possible production scheduling combinations based on material length, quantity, and width; Normalize each candidate production scheduling combination and calculate the Euclidean norm with actual demand: If two candidate production scheduling combinations Combined with candidate production schedule The production scheduling is considered successful if the following conditions are met: ; in, Candidate scheduling combination The nth cutting group in Candidate scheduling combination The nth cutting group in For order length information, symbol In order to normalize the three-dimensional feature vector of the candidate production scheduling combination, the symbol is the Euclidean norm calculation, The difference vector between the normalized order demand feature vector and the normalized feature vector of the candidate scheduling combination plan C2; Set similarity threshold : ; in, is the dynamic adjustment factor, ; If the similarity threshold Greater than probability similarity , then the two candidate production scheduling combinations are eliminated, otherwise the two candidate production scheduling combinations are retained; Based on the error range of steel length, scatter fitting regression calculation is performed to obtain the optimal production scheduling plan: Perform scatter point fitting on the set of candidate production schedules after screening, taking into account the 10% length error range of the steel after cutting, to obtain a scatter point fitting set;

[0042] in, =Theoretical length of steel ordered , = Actual cutting length deviation rate ; Length error range constraints: ; Use the least squares method to determine the optimal production scheduling plan, minimizing the error between the actual production scheduling results and the ideal production scheduling plan. Specifically, this includes: a) Construct the input matrix D, which contains the theoretical length of steel for the order Deviation rate from actual cutting length ; b) solving the weighted least squares equation to obtain the deviation prediction model; c) Select the candidate production scheduling sequence that minimizes the absolute value of the forecast deviation as the optimal production scheduling plan.

[0043] The calculation formula of the length deviation rate is as follows: Length deviation rate ;in, The length of steel required for the order, The actual cutting length.

[0044] S34 reduces the simulated annealing temperature and repeats the above S32 to S33 until the termination condition is met, and finally determines the optimal production scheduling plan; if the requirements are not met, returns to S31 to readjust the parameters and iterate.

[0045] Application Examples This example is based on a heavy equipment manufacturing company (with an annual production capacity of 500,000 tons). The company's original production scheduling system suffered from low material utilization (72%), high delay rate for urgent orders (18%), and uneven equipment load (Gini coefficient 0.48). The implementation process after adopting the solution of the present invention is as follows: S1 data matrix construction There are currently 50 orders to be queued (including 10 urgent orders), involving 5 materials such as Q235B and Q355C; Parameters of the three cutting machines: maximum processing length 12m, real-time task queue lengths 8 / 15 / 5 respectively; 20 rolls of coils in stock (length 1.8-12.3m); The matrix is ​​constructed as follows: Order feature matrix

[0046]

[0047] Device Capability Matrix

[0048]

[0049] Residual Inventory Matrix

[0050]

[0051] Dynamic calculation of emergency order penalty coefficient: = 0.5 + 0.1 × (number of emergency orders / total number of orders) = 0.5 + 0.1 × (10 / 50) = 0.7.

[0052] S2 Dynamic Risk Assessment Monte Carlo Sequence Generation: In Constrained Space Generate 200 initial sequences within ), after device capability verification (e.g., DevB maximum queue length 12), illegal sequences are eliminated and 120 candidate sequences are retained.

[0053] Scrap optimization example: Traditional solution: Order Ord01 (requires 6200mm) and use Coil01 (8500mm) alone Loss rate = (8500 - 6200) / 8500 × 100% = 27.1%.

[0054] This solution combines cutting: Ord01 (6200mm) + Ord12 (1900mm) using Coil01 Loss rate = (8,500 - (6,200 + 1,900)) / 8,500 × 100% = 9.8% (a 64% reduction).

[0055] Delay risk prediction: DevB load rate = current queue / maximum capacity = 15 / 12 = 125% Estimated start time = 37.5 minutes delay Order Ord01 delay risk = 22.5min Comprehensive loss value 9.8% + 0.7 × log(1 + 22.5 / 60)≈15.3%.

[0056] S3 Iterative Optimization Simulated annealing parameters: Initial temperature = 1000, attenuation coefficient α = 0.95, 50 iterations Adaptive loss threshold = 0.9 × 28% = 25.2%, tightening by 5% per iteration; Area Crossing Operation: Fixed priority ≥ 4 orders (such as Ord01) Perform crossover for orders with priority ≤ 3: randomly select the [3,15] order interval to swap.

[0057] Regression fit verification: Length error dataset

[0058] Least squares fitting results: the absolute error is reduced by 72%.

[0059] Comparison of implementation effects (6-month average)

[0060] The method of the present invention obtains order parameters, equipment status and inventory balance through a real-time data interface, and converts them into a multidimensional numerical matrix (including an order feature matrix, an equipment capability matrix, and a surplus inventory matrix) to ensure structured processing and efficient use of data; adopts the Monte Carlo method to integrate multiple strategies to optimize the production scheduling process, generates candidate sequences based on order priority and equipment constraints, dynamically evaluates scrap loss and delay risks, combines simulated annealing and an adaptive threshold mechanism to achieve sequence crossover and mutation, and determines the optimal production scheduling plan within the allowable error range through similarity screening and regression fitting. It iterates repeatedly until the conditions are met, thereby balancing global optimization and local efficiency, and improving production fault tolerance and resource utilization.

[0061] Based on this, the present invention further proposes a steel cutting production scheduling device based on dynamic risk assessment, which is used to execute the above-mentioned steel cutting production scheduling method based on dynamic risk assessment, including: The data acquisition and processing module acquires production scheduling data in real time. The production scheduling data includes order parameter sets, equipment state vectors, and inventory balances. The module then performs numerical processing on the production scheduling data, converting it into a multidimensional numerical matrix. The multidimensional numerical matrix includes an order feature matrix, an equipment capability matrix, and a surplus inventory matrix. This data acquisition and processing is used to execute step S1 of the aforementioned method for steel cutting scheduling based on dynamic risk assessment.

[0062] The dynamic risk assessment module, based on the Monte Carlo method, operates on the order feature matrix, the equipment capacity matrix, and the surplus inventory matrix within the constraint space to generate a candidate production schedule that satisfies order priorities and equipment constraints. For each order in the candidate production schedule, it dynamically calculates scrap loss and delay risk to comprehensively determine a comprehensive loss assessment for each order. The dynamic risk assessment module is used to execute step S2 of the aforementioned steel cutting production scheduling method based on dynamic risk assessment.

[0063] The production scheduling generation module, combining a simulated annealing algorithm with an adaptive loss threshold mechanism, performs crossover and mutation operations on the candidate production scheduling sequences. It then eliminates redundant sequences through similarity screening and uses regression fitting to determine the optimal potential production scheduling within the tolerance range. The sequence crossover, mutation, screening, and fitting process are iteratively executed until the optimal production scheduling solution that meets the preset conditions is obtained. The production scheduling generation module is used to execute step S3 of the aforementioned steel cutting scheduling method based on dynamic risk assessment.

[0064] This invention achieves highly adaptable production scheduling by constructing a matrix processing model for multi-source data, combining Monte Carlo random sampling with a dynamic loss rate feedback mechanism. The algorithm innovatively employs a probabilistic component similarity matching method to generate candidate sequences based on order length and width parameters. Through real-time loss rate calculation and a dynamic threshold optimization mechanism, it performs a global search within the solution space, generating an optimal production scheduling plan that balances material utilization and production feasibility.

[0065] The present invention further provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device.

[0066] The computer-readable medium carries one or more programs. When the one or more programs are executed by an electronic device, the electronic device implements the method in the above embodiment.

[0067] It should be noted that although several modules or units of the device for action execution are mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0068] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0069] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein.

[0070] The above is only a specific implementation of the present invention, but the design concept of the present invention is not limited to this. Any non-substantial changes to the present invention using this concept shall be deemed as an infringement of the protection scope of the present invention.

Claims

1. A steel cutting production scheduling method based on dynamic risk assessment, characterized in that: The steps include: S1 acquires production scheduling data in real time, the production scheduling data including an order parameter set, an equipment state vector, and an inventory balance; performs numerical processing on the production scheduling data and converts it into a multidimensional numerical matrix, the multidimensional numerical matrix including an order feature matrix, an equipment capacity matrix, and a surplus inventory matrix; S2 uses the Monte Carlo method to calculate the order feature matrix, the equipment capacity matrix, and the surplus inventory matrix within the constraint space to generate a candidate production schedule that satisfies the order priority and equipment constraints. For each order in the candidate production schedule, it dynamically calculates scrap loss and delay risk to comprehensively obtain a comprehensive loss assessment value for each order. S3 combines the simulated annealing algorithm and the adaptive loss threshold mechanism to perform crossover and mutation operations on the candidate production scheduling sequences; eliminates redundant sequences through similarity screening, and uses regression fitting to determine the potential optimal production scheduling plan within the allowable error range; and iteratively executes the sequence crossover, mutation, screening and fitting process until the optimal production scheduling plan that meets the preset conditions is obtained.

2. A steel cutting production scheduling method based on dynamic risk assessment according to claim 1, characterized in that: The production scheduling data includes the order parameter set , device state vector , inventory balance distribution and material replacement rules table ;in, is the total number of orders currently to be scheduled. For the An order data object; is the total number of available production equipment, For the Status data of each device; is the total number of coil types in stock, For the Inventory data for coil-like materials; extract the geometric parameters of the order based on the order parameter set, including the length, width, thickness, and quantity of the steel required for the order, mark the order priority weight, and dynamically adjust it based on delivery urgency and customer level. The order feature matrix constructed is as follows: ； in, is the order number, The length of steel required for the order, The width of steel required for the order, The quantity of steel required for the order, is the priority weight; Based on the device state vector, the real-time throughput, current task load, and maintenance status of the device are obtained. The wear index of the device is calculated and the remaining service life is predicted by combining historical data. The device capability matrix is ​​constructed as follows: ； in, is the device number, is the order length, is the maximum processing length, is the current task queue length; Based on the inventory balance distribution, the remaining size, material type, and storage location of the coils in each warehouse are recorded. The recommended usage scenarios and replacement rules of the remaining materials are marked. The remaining material inventory matrix is ​​constructed as follows: ； in, The coil number, is the ascending length, The number of the replaceable material.

3. The steel cutting production scheduling method based on dynamic risk assessment according to claim 1, characterized in that: Step S2 specifically includes the following: S21 sets constraint space parameters, including order priority sorting rules, equipment operation thresholds, and production resource constraints; performs random sampling on the order feature matrix, equipment capability matrix, and surplus inventory matrix based on the Monte Carlo method to generate an initial production scheduling sequence within the constraint space; S22 performs compliance verification on the initial production scheduling sequence based on the order priority sorting rules and equipment operation thresholds, eliminates sequences that do not meet the constraints, and retains candidate production scheduling sequences that meet the requirements; S23 calculates the scrap loss value for each order in the candidate production scheduling sequence by combining the material demand parameters in the residual material inventory matrix and the order feature matrix; Calculate the delay risk coefficient of each order based on the equipment load parameters in the equipment capacity matrix and the delivery time parameters in the order feature matrix; S24 combines the scrap loss value and the delay risk coefficient according to the preset weight to obtain the comprehensive loss assessment value of each order.

4. The steel cutting production scheduling method based on dynamic risk assessment according to claim 3, characterized in that: In step S21, the Monte Carlo search space is defined as follows: ； in, represents the Monte Carlo search space, i.e., the set of candidate production schedules. Indicates the Candidate scheduling sequences, represents the full permutation of the order set, Indicates the index of the currently generated sequence, Indicates the maximum number of generated sequences.

5. The steel cutting production scheduling method based on dynamic risk assessment according to claim 3, characterized in that: In step S24, the comprehensive loss assessment value is calculated as follows: From the rest stock matrix Filter the coils that meet the conditions: ; in, For a collection of coils that meet the conditions, For the A roll of material, is the remaining length of the coil, The replaceable material number of the coil, The length of steel required for the order, The material type of the order. It is the material replacement rule function; Select the optimal coil material through the value function: ; in, is the optimal coil material selection function, In order to find the coil object with the minimum comprehensive cost in the candidate list, For size matching, is the transportation distance from the warehouse to the processing equipment, To calculate transportation costs; For successfully matched orders , calculate the scrap loss value: ； in, The length of the selected coil, It is The order length of the sub-cut, The number of orders cut from the coil; Processing start time prediction: ； in, is the current system time, is the device queue load rate, For equipment Real-time throughput, For equipment The length of the task queue, is the unit time basis; Delay risk factor calculation: ； in, The processing time of the order, The delivery deadline for the order; Calculate the The comprehensive loss evaluation value of the iteration : ； in, For material consumption, is the time consumption item, is the delay penalty coefficient; ； in, Indicates the number of urgent orders, The total number of orders.

6. The steel cutting production scheduling method based on dynamic risk assessment according to claim 1, characterized in that: Step S3 specifically includes the following: S31 initializes simulated annealing parameters, including initial temperature, cooling coefficient and termination temperature, and sets an initial reference value of the adaptive loss threshold; S32 selects the candidate production scheduling sequence with the best comprehensive loss assessment value as the parent generation, fixes the position of the high-priority order, and performs a regional crossover operation on the remaining orders to generate a child sequence; the child sequence obtained by regional crossover is randomly swapped to generate a new candidate production scheduling sequence; S33 combines the adaptive loss threshold mechanism with the simulated annealing algorithm to evaluate and accept the new candidate production scheduling sequence, and screens the candidate production scheduling sequences through probabilistic similarity calculation to retain the sequences that meet the requirements; based on the steel length error range, a scatter point fitting regression calculation is performed on the screened sequences to determine the potential optimal production scheduling plan; S34 reduces the simulated annealing temperature and repeats the above S32 to S33 until the termination condition is met, and finally determines the optimal production scheduling plan; if the requirements are not met, returns to S31 to readjust the parameters and iterate.

7. The steel cutting production scheduling method based on dynamic risk assessment according to claim 6, characterized in that: Step S32 includes: Take two candidate production scheduling sequences with better comprehensive loss assessment values ​​as parent sequences, fix the positions of orders with priority ≥ 4, and perform region intersection and random position exchange on the remaining orders to generate subsequences as new candidate production scheduling sequences: Region crossover operation: Randomly select the crossover interval [start, end] from the two parent sequences, where start is the starting position order of the parent sequence and end is the ending position order of the parent sequence. The child sequence directly inherits the interval fragment of one of the parent sequences, and the remaining positions are filled with orders with priority ≤ 3 according to the order of the other parent sequence; Random position swap: Perform n swaps on each subsequence, and only swap the positions of orders with priority ≤ 3, n = 1, 2, 3, ... 5; Combined with the simulated annealing mechanism to accept suboptimal solutions with probability, specifically: Annealing temperature update formula: ； in, For the The annealing temperature at the iteration, is the initial temperature of the scheduling process, is the decay rate, is the number of iterations; The probability of accepting a suboptimal solution uses the Metropolis criterion: ； in, is the probability of accepting a suboptimal solution, is the difference between the loss evaluation values ​​of the new solution and the current solution, For the Annealing temperature at the iteration; Set the adaptive loss threshold and gradually tighten the standard during iteration, specifically: Initial loss threshold ,in is the historical average loss rate; Loss threshold varies with the number of iterations Adaptive Tightening: ， ； Set dynamic acceptance probability, allowing the algorithm to accept inferior solutions with a certain probability ( > ), avoid falling into the local optimum, gradually reduce the probability of accepting inferior solutions as the iteration proceeds, and finally stabilize at the optimal solution: ;in, is the comprehensive loss value of the current iteration, is the historical optimal loss value, temperature parameter With the number of iterations attenuation: , is the initial temperature.

8. The steel cutting production scheduling method based on dynamic risk assessment according to claim 7, characterized in that: In step S33, candidate production scheduling sequences are screened by probability similarity, specifically including: Collect order information in the candidate scheduling sequence, including material width, length, and quantity; Generate all possible production scheduling combinations based on material length, quantity, and width; Normalize each candidate production scheduling combination and calculate the Euclidean norm with actual demand: If two candidate production scheduling combinations Combined with candidate production schedule The production scheduling is considered successful if the following conditions are met: ； in, Candidate scheduling combination The nth cutting group in Candidate scheduling combination The nth cutting group in For order length information, symbol In order to normalize the three-dimensional feature vector of the candidate production scheduling combination, the symbol is the Euclidean norm calculation, The difference vector between the normalized order demand feature vector and the normalized feature vector of the candidate scheduling combination plan C2; Setting similarity threshold : ； in, is the dynamic adjustment factor, ; If the similarity threshold Greater than probability similarity , then the two candidate production scheduling combinations are eliminated, otherwise the two candidate production scheduling combinations are retained; Based on the error range of steel length, scatter fitting regression calculation is performed to obtain the optimal production scheduling plan: Perform scatter point fitting on the set of candidate production schedules after screening, taking into account the 10% length error range of the steel after cutting, to obtain a scatter point fitting set; ； in, =Theoretical length of steel ordered , = Actual cutting length deviation rate ; Length error range constraints: ； Use the least squares method to determine the optimal production scheduling plan, minimizing the error between the actual production scheduling results and the ideal production scheduling plan. Specifically, this includes: a) Construct the input matrix D, which contains the theoretical length of steel for the order Deviation rate from actual cutting length ; b) Solve the weighted least squares equation to obtain the deviation prediction model; c) Select the candidate production scheduling sequence that minimizes the absolute value of the prediction deviation as the optimal production scheduling plan.

9. The steel cutting production scheduling method based on dynamic risk assessment according to claim 8, characterized in that: The calculation formula of the length deviation rate is as follows: Length deviation rate ; in, The length of steel required for the order, The actual cutting length.

10. A steel cutting and production scheduling device based on dynamic risk assessment, characterized in that: include: A data acquisition and processing module acquires production scheduling data in real time, wherein the production scheduling data includes order parameter sets, equipment state vectors, and inventory balances; Performing numerical processing on the production scheduling data and converting it into a multidimensional numerical matrix, wherein the multidimensional numerical matrix includes an order feature matrix, an equipment capacity matrix, and a surplus material inventory matrix; The dynamic risk assessment module, based on the Monte Carlo method, operates on the order feature matrix, the equipment capacity matrix, and the surplus inventory matrix within the constraint space to generate a candidate production schedule that satisfies order priorities and equipment constraints. For each order in the candidate production schedule, it dynamically calculates scrap loss and delay risk to comprehensively determine a comprehensive loss assessment value for each order. The production scheduling plan generation module combines the simulated annealing algorithm and the adaptive loss threshold mechanism to perform crossover and mutation operations on the candidate production scheduling sequences; eliminates redundant sequences through similarity screening, and uses regression fitting to determine the potential optimal production scheduling plan within the allowable error range; and iteratively executes the sequence crossover, mutation, screening and fitting process until the optimal production scheduling plan that meets the preset conditions is obtained.

Images