Multi-base, multi-production-line and multi-process global plan optimization method and system for steel production
By introducing a global coordination and optimization mechanism into steel production, the problem of global imbalance in modular production planning was solved, and overall optimization of multiple bases, production lines, and processes was achieved. This improved the feasibility and stability of production plans, reduced costs, and enhanced adaptability to production disturbances.
Patent Information
- Application Number
- CN202610326109.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-17
- Publication Date
- 2026-05-12
AI Technical Summary
The lack of a unified overall coordination mechanism in steel production leads to local optima in modular production planning but global imbalance, making it difficult to achieve overall optimization across multiple bases, production lines, and processes.
A global coordination and optimization mechanism is introduced to transfer data during the module connection process. Through global evaluation and feedback optimization, a production plan scheme that meets global constraints is generated, including steps such as order standardization, hierarchical clustering, surplus material matching, multi-dimensional feature analysis, and cross-process batching.
It enables global collaborative optimization of production plans across multiple bases, production lines, and processes, improving the feasibility and stability of production plans, increasing resource utilization efficiency, reducing costs, and enhancing adaptability to production disturbances.
Smart Images

Figure CN122022380A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of steel production planning and optimization control technology, specifically relating to a global planning optimization method and system for multi-base, multi-production line, and multi-process steel production. Background Technology
[0002] Steel production is a typical process industry, spanning multiple tightly coupled processes such as steelmaking, continuous casting, rolling, and post-processing. Furthermore, in the context of group-based production, companies often possess multiple production bases and production lines. As market demand shifts from large-batch, single-variety production to small-batch, multi-variety production with more dispersed delivery times, the complexity of steel companies' production organization has significantly increased.
[0003] In actual production management, steel companies typically employ modular production planning to address the complexity of production under conditions of multiple bases, production lines, and processes. This type of production planning system generally divides functions such as order processing, multi-base order allocation, utilization of surplus inventory, multi-production line order configuration, and multi-process batching into several relatively independent business modules, which operate sequentially or in parallel according to predetermined business processes. Through this modular approach, companies can manage and optimize specific aspects such as order organization, inventory utilization, production line load allocation, and process coordination. For example, order batching can increase the batch size per production run, surplus inventory can reduce new production demand, and production line allocation can alleviate the load pressure on individual production lines, thereby improving production organization efficiency to a certain extent.
[0004] However, current technology still has the following shortcomings: Currently, steel companies often suffer from fragmented production planning, lacking a unified coordination mechanism. Each functional module typically operates sequentially or independently according to fixed business processes. Its optimization goals and constraints are limited to the module itself, making it difficult to perceive the feasibility of downstream modules and the overall impact, easily leading to a situation of "local optima at the expense of global balance."
[0005] Existing technologies lack global optimization for the data transfer process between modules. Most existing solutions evaluate or make simple corrections to the results after the module is completed, but fail to introduce a unified global evaluation and feedback mechanism during the data transfer process between modules. As a result, once the preceding decisions are determined, subsequent modules can only passively accept them and find it difficult to make dynamic adjustments.
[0006] In multi-site scenarios, order allocation, surplus material utilization, production line load, and cross-process rhythm are highly coupled. Without a unified global coordination strategy, problems such as capacity imbalance between sites, poor process connection, and decreased equipment utilization are likely to occur.
[0007] Therefore, how to introduce a global coordination and optimization mechanism that runs through the module-connected process while maintaining the existing modular production planning system, so as to achieve the overall optimization of production plans for multiple bases, multiple production lines and multiple processes, has become a technical problem that urgently needs to be solved in the field of steel production planning. Summary of the Invention
[0008] In order to overcome the above-mentioned problems in the existing technology, the present invention provides a global planning optimization method and system for multi-base, multi-production line, and multi-process steel production, which is applicable to steel enterprises to carry out collaborative production organization and global planning decision-making for small-batch, multi-variety orders under the conditions of multiple bases, multiple production lines, and multiple processes.
[0009] A global planning optimization method for multi-site, multi-production line, and multi-process steel production, the method comprising: S1. Receive sales order data and perform order standardization processing to obtain an order set. Merge or split the order set to form a set of orders to be allocated. S2. Perform hierarchical clustering and base matching decisions on the set of orders to be assigned to obtain a base-level order set and its base affiliation; S3. Match the base-level order set with the inventory surplus material set in terms of specifications, materials and / or dimensions to generate a surplus material solution, and deduct and mark the net production demand of the base-level order set according to this solution to obtain the base-level order set after surplus material adjustment; S4. Perform multidimensional feature analysis and dynamic clustering on the base-level order set after adjusting the surplus materials within each base, and configure the order clusters formed by the clustering results to specific production lines to obtain production line-level order sets; S5. Based on the process constraints and economic batch requirements of steelmaking-continuous casting-rolling-post-processing, the production line-level order set is batched across processes to generate the batch structure and connection relationship of furnace batch, casting batch, rolling batch or post-processing batch, and obtain the cross-process batch plan; S6. Set up a global coordination optimization model, calculate global evaluation indexes for the data transfer process between steps S2 and S5, and output parameter correction instructions to adjust the parameters or constraints of the previous or next step. S7. Based on the cross-process batch plan and the set of parameters or constraints after global coordination and optimization, generate a production plan scheme that can be optimized and updated for the planning cycle.
[0010] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the order set includes steel type, specifications, delivery date, quantity and special process requirements, and the merging or splitting of the order set specifically includes: the order merging includes: clustering and merging orders with the same or compatible steel type, similar specifications and delivery dates within a preset time window to form candidate production batches; The order splitting includes: for ultra-large batch orders or orders with delivery time spans exceeding a threshold, splitting the order into multiple sub-orders based on base process boundaries, capacity limitations, or delivery time segments, so that each sub-order can be entered into the base allocation decision.
[0011] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the hierarchical clustering and base matching decision is based on the capacity structure, process adaptability range, transportation and manufacturing costs, base inventory and production status of each base. When making decisions, the differences in manufacturing costs of each base, cross-base transportation costs, inventory holding costs, order and base process matching degree, base key process load level and load balance constraints are comprehensively considered to obtain the allocation result structure data that includes the impact of order-base-expected cost-base load.
[0012] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the matching specifically includes: calculating the substitutable ratio of surplus materials to orders, generating surplus materials to act as orders; marking the orders to be acted as with a result identifier, and passing the mark result identifier along with the order data to step S4.
[0013] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein S4 specifically includes: constructing an order feature vector based on the product category, specifications, material requirements, delivery constraints, and special process requirements of the order; under the premise of satisfying the production line process adaptation range constraints, dynamically clustering the order feature vector to form a candidate order cluster, and configuring the candidate order cluster to the target production line according to the production line processing capacity, current load level, and energy consumption characteristics, and outputting a production line-level order set.
[0014] In addition to the aspects and any possible implementations described above, an implementation is further provided, wherein S5 specifically includes: organizing furnace batches according to steel type and furnace capacity in the steelmaking stage; cascading the furnace batches in the continuous casting stage according to the casting sequence; organizing rolling batches according to billet specifications and mill roll changing constraints in the rolling stage; and organizing post-processing batches according to customer specifications and surplus material utilization strategies in the post-processing stage, and generating cross-process connection constraint data that reflects the quantity and sequence matching relationship between processes.
[0015] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein S6 specifically includes: first calculating a global evaluation index set including at least delivery date achievement rate, capacity utilization rate, equipment utilization rate, capacity balance, energy consumption level and inventory level; then generating a parameter correction instruction based on the index set; the parameter correction instruction is used to adjust the clustering threshold, weight coefficient, constraint enable / disable status or allocation priority, so that the output of subsequent modules satisfies the global constraint consistency.
[0016] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the production planning scheme generation includes: calculating the load utilization rate of each process at each base based on cross-process batch planning, identifying the overload risk of bottleneck processes or key equipment; using capacity balance constraints and equipment utilization rate constraints as constraints or optimization objectives for global optimization, and adopting a cross-process hierarchical planning coordination method to transmit constraints and adjust the pace step by step, so that the steelmaking pace and continuous casting times, continuous casting billet supply and rolling start sequence, and post-processing capacity meet an executable spatiotemporal matching relationship.
[0017] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the optimization update includes: during the planning period, when order insertion, equipment failure, maintenance plan changes, or surplus material inventory updates cause changes in constraints, dynamically updating the global constraint set and triggering a second solution; the solution employs mathematical programming, intelligent optimization, or a combination thereof, and when multiple objectives conflict, it uses weighted summation, hierarchical optimization, or Pareto strategy to output an updated production plan.
[0018] This invention also provides a global planning optimization system for multi-site, multi-production line, and multi-process steel production, the system being used to implement the method, comprising: The order and demand management module is used to receive sales order data and perform order standardization processing to obtain an order set, and to merge or split the order set to form a set of orders to be allocated. The multi-base order hierarchical processing and allocation module is used to perform hierarchical clustering and base matching decisions on the set of orders to be allocated, so as to obtain a base-level order set and its base affiliation; The inventory surplus material matching module is used to match the base-level order set with the inventory surplus material set in terms of specifications, materials and / or dimensions, generate a surplus material solution, and deduct and mark the net production demand of the base-level order set according to the solution to obtain the base-level order set after surplus material adjustment. The multi-production-line order allocation module is used to perform multi-dimensional feature analysis and dynamic clustering on the base-level order set after adjusting the surplus materials within each base, and to configure the order clusters formed by the clustering results to specific production lines to obtain production-line-level order sets. The multi-process collaborative batching module is used to perform cross-process joint batching of the production line-level order set based on the process constraints and economic batch requirements of steelmaking-continuous casting-rolling-post-processing, generating the batch structure and connection relationship of furnace batch, casting batch, rolling batch or post-processing batch, and obtaining the cross-process batch plan; The global coordination and optimization module is used to calculate global evaluation indicators and output parameter correction instructions for the data transfer process between the multi-base order hierarchical processing and allocation module, the inventory surplus material matching module, the multi-production line order allocation module and the multi-process collaborative batching module, so as to adjust the parameters or constraints of the previous or next step. The plan generation and rolling optimization module is used to generate a production plan scheme that can be optimized and updated for the planning cycle based on the cross-process batch plan and the set of parameters or constraints after global coordination and optimization.
[0019] Beneficial effects of the present invention This invention proposes a multi-base, multi-production-line, and multi-process global planning optimization method and system that introduces global evaluation and feedback optimization during the data transmission process of module concatenation. Modules operate sequentially according to business processes, and global coordination optimization is performed during data transmission between modules, ultimately generating a production plan scheme that meets global constraints. The global optimization mechanism is embedded in the data flow of module concatenation, rather than as a post-event evaluation step, thereby guiding each module to tend towards the global optimum through dynamic adjustments. Through this method, in the sequential processing of orders from "base allocation—surplus material matching—production line allocation—multi-process batching," a global perspective is continuously introduced to evaluate intermediate results, and the evaluation results are applied back to module parameters and constraints. This ensures the independence of each module while achieving collaborative optimization of the production plan at the global level. The beneficial technical effects of this invention are as follows: (1) Achieving global collaborative optimization of production plans across multiple bases, production lines, and processes: This invention connects functional modules such as order processing, multi-base allocation, in-stock material matching, multi-production line allocation, and multi-process collaborative batching according to business processes, and introduces a global evaluation and coordination optimization mechanism during data transfer between modules. This allows the production plan to be continuously constrained and guided from a global perspective during its generation, avoiding the local optimum problem caused by the individual optimization of each module in traditional modular solutions, and achieving global collaborative optimization at the multi-base, multi-production line, and multi-process levels. Explainable intelligent monitoring: The large language model outputs diagnostic results while providing diagnostic reasons and operational suggestions in natural language form, making it easier for remote control room operators to understand the alarm basis, improving their trust in the intelligent system and the efficiency of fault handling; (2) Introducing a global feedback mechanism while maintaining module independence to improve system scalability: This invention does not change the original business responsibilities and operating methods of each functional module, but instead introduces global evaluation results into the data flow of the module series to provide feedback and adjust the parameters and constraints of subsequent modules, thereby achieving global coordination and optimization. This approach is compatible with the modular architecture of existing production planning systems, facilitates integration with existing ERP, MES, or APS systems, and has good engineering scalability and system adaptability.
[0020] (3) Improve the overall executability and stability of production plan results: By comprehensively considering multi-dimensional indicators such as delivery date achievement rate, capacity utilization and balance, energy consumption level and inventory turnover during the production plan generation process, and evaluating and constraining the fluctuation of module output results under different loads and different production scenarios, this invention can effectively reduce the drastic fluctuations of production plan in the rolling optimization process, improve the stability and continuous performance of production plan in complex scenarios, and enhance the executability of plan results.
[0021] (4) Improve the efficiency of production resource utilization and reduce overall operating costs: In the process of global evaluation and coordination optimization, this invention incorporates the capacity utilization and equipment utilization rate of each base, production line and process into a unified evaluation system. Through global coordination, it guides order allocation, production line configuration and process batching decisions, effectively alleviates the imbalance of base capacity and the bottleneck of key equipment, improves the overall efficiency of production resource utilization, and thus reduces manufacturing costs, energy costs and inventory costs.
[0022] (5) Support dynamic rolling optimization and enhance the adaptability to production disturbances: When production disturbances such as order changes, equipment status changes, maintenance plan adjustments or inventory surplus updates occur, the present invention can dynamically update the constraints and model parameters based on the global evaluation coordination optimization model, and re-execute the module serial process to output the updated production plan, thereby improving the system's responsiveness and robustness to uncertain production environments.
[0023] (6) Applicable to multi-variety, small-batch steel production scenarios, with significant engineering application value: This invention is geared towards the production characteristics of multi-variety, small-batch, and dispersed delivery times in the steel industry. By introducing a global coordination mechanism in order organization, surplus material utilization, and cross-process batching, it can improve order concentration and reduce losses caused by frequent switching, effectively alleviating the problems of high cost and difficulty in organizing small-batch production, and has good engineering application prospects. Attached Figure Description
[0024] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation
[0025] To better understand the technical solution of this invention, the content of this invention includes, but is not limited to, the specific embodiments described below. Similar technologies and methods should be considered within the scope of protection of this invention. To make the technical problem to be solved, the technical solution, and the advantages of this invention clearer, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments.
[0026] It should be understood that the embodiments described in this invention are merely some embodiments of this invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.
[0027] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.
[0028] like Figure 1 As shown, this invention provides a global planning optimization method for multi-site, multi-production line, and multi-process steel production, the method comprising: S1: Input sales orders and customer demand information. By constructing a multi-base, multi-production line, and multi-process collaborative production model, the input data is standardized. The processed order set serves as the basic input data for subsequent multi-base allocation. S2: After completing the order standardization process, perform multi-base order tiered processing; this process comprehensively considers the differences in manufacturing costs, transportation costs, and inventory costs of each production base, as well as the base's process adaptability for specific steel grades and specifications and the current production line load. S3: After obtaining the base-level order set, perform parallel matching of surplus inventory materials. This module, based on a rule knowledge base and matching algorithm, compares the order requirements with the specifications, materials, and dimensions of the surplus inventory materials to generate a solution using the surplus materials.
[0029] S4: Within a single production base, the adjusted order set is allocated across multiple production lines. This method divides highly adaptable orders into order clusters and assigns them to corresponding production lines, thereby forming a stable and executable production line-level order set, providing clear boundaries for multi-process collaborative batching.
[0030] S5: Based on the production line-level order set, perform multi-process collaborative batching. According to the process constraints and economic batch requirements of each process in steelmaking, continuous casting, rolling and post-processing, perform cross-process joint batching of orders.
[0031] S6: In the process of multi-base allocation, surplus material matching, multi-production line allocation, and multi-process batching, this invention introduces a global coordination and optimization mechanism. At key data transmission nodes, a global evaluation model is constructed. The evaluation results are not directly used to select solutions, but rather to generate parameter correction instructions to adjust subsequent clustering thresholds, weight coefficients, constraints, or allocation priorities.
[0032] S7: After completing cross-process collaborative batching and global coordination optimization, a production plan scheme oriented towards the planning cycle is generated. This scheme includes the production arrangements and load distribution of each base, production line, and process. The generation of the production plan scheme includes: calculating the load utilization rate of each process at each base based on the cross-process batch plan, identifying the overload risk of bottleneck processes or key equipment; using capacity balance constraints and equipment utilization rate constraints as constraints or optimization objectives for global optimization, and adopting a cross-process hierarchical planning coordination method to transmit constraints and adjust the pace of adjustment step by step, so that the steelmaking pace and continuous casting times, continuous casting billet supply and rolling start sequence, and post-processing capacity meet an executable spatiotemporal matching relationship.
[0033] Further, in step S1, order processing includes: structuring the steel type, specifications, quantity, delivery date, and special process requirements of the orders, and eliminating differences in expression format and granularity between orders from different sources. Based on this, order merging and splitting operations are performed according to production organization needs. Specifically, for orders with small batches, similar specifications, and close delivery windows, a hierarchical clustering method is used to merge them into candidate production batches to improve production concentration and single-batch production efficiency; for orders with excessively large batches or large delivery spans, the orders are split according to process capacity boundaries and capacity constraints, allowing them to be allocated to different production bases or different time windows for execution. The order set processed in the above way serves as the basic input data for subsequent multi-base allocation.
[0034] Furthermore, in step S2, information such as the capacity structure, process adaptability range, manufacturing cost, transportation cost, and inventory cost differences of each production base, as well as the base's process adaptability to specific steel grades and specifications, and the current production line load level, are uniformly incorporated into the decision-making model. Through a multi-objective optimization strategy, the optimal base allocation for each order or order group is determined under the premise of satisfying order delivery time and process constraints.
[0035] Further, in step S3, the base-level order set output by the multi-base order hierarchical processing and allocation module and the current inventory surplus information are used as input. A rule knowledge base and matching algorithm are invoked to perform constraint judgment and suitability calculation on order requirements and inventory surplus in dimensions such as steel grade, chemical composition range, specification and size parameters, quality grade, and processable path. The rule knowledge base provides hard constraint judgment rules, including steel grade substitution rules, specification and size processable boundary rules, customer disallowed rules, quality grade matching rules, and process feasibility rules. Suitability calculation only proceeds when the inventory surplus meets all the above hard constraint conditions. The matching algorithm scores the order-surplus combination that meets the constraints using multiple indicators, calculates a comprehensive matching score, and generates surplus as the decision result under the premise of satisfying inventory quantity constraints and order requirement constraints. The decision result includes: a. The quantity of surplus materials corresponding to the order; b. The remaining demand of the order; c. Information on the source of surplus materials, including the base, warehouse location and batch identification; d. The proportion of surplus materials used as inventory.
[0036] When surplus inventory can meet all or part of the order demand, the system automatically generates a surplus material solution and deducts the original order demand to form the net production demand. When there is no surplus inventory that meets the constraints, the original order demand remains unchanged. The order data adjusted by surplus material matching will be updated to a data structure containing the "net production demand" field and transmitted to the multi-production line order allocation module along with the order's optimal base allocation information. This ensures that subsequent production line load calculations, production line capacity matching, and process batching decisions are all based on the net production demand, thereby guaranteeing that production resource allocation is consistent with actual production demand.
[0037] Further, in step S4, an order feature vector is constructed based on the product category, specifications, material requirements, delivery constraints, and special process requirements of the order. Simultaneously, dynamic clustering analysis is performed on the order feature vector by considering the processing capacity limits, process adaptability range, current load level, and energy consumption characteristics of each production line. Through this process, orders with high adaptability and similar production conditions are grouped into candidate order clusters, and these order clusters are assigned to their corresponding production lines, thus forming a clearly structured and load-controllable production line-level order set. This result provides clear production line boundaries and order inputs for subsequent multi-process collaborative batching.
[0038] Furthermore, in step S5, during the steelmaking stage, furnace batches are organized according to steel grade and furnace capacity; during the continuous casting stage, multiple furnace batches are connected in series into a casting sequence according to the casting order; during the rolling stage, rolling batches are organized according to billet specifications, mill capacity, and roll changing requirements; and during the post-processing stage, processing batches are arranged in conjunction with customer specifications and surplus material utilization strategies. Through the above-mentioned coordinated batching, a batch structure with smooth cross-process connections and matching quantity and sequence is formed.
[0039] Further, in step S6, the global coordination optimization includes: calculating a global evaluation index set based on the intermediate data transmitted between modules, wherein the index set includes at least delivery date achievement rate, capacity utilization rate, equipment utilization rate, capacity balance, energy consumption level, and inventory level; generating parameter correction instructions based on the index set, wherein the parameter correction instructions are used to adjust the clustering threshold, weight coefficient, constraint enable / disable status, or allocation priority of at least one module, so that the output of subsequent modules satisfies global constraint consistency.
[0040] Furthermore, the rolling optimization includes: during the planning period, when disturbance events such as order insertion, equipment failure, maintenance plan changes, or surplus material inventory updates occur, causing changes in production capacity constraints, inventory constraints, or delivery date constraints, the system automatically updates the global constraint set and triggers a re-solution of the production planning model.
[0041] The re-solution process includes: 1) Construct a multi-objective optimization model that includes cost targets, delivery date achievement rate targets, capacity utilization rate targets, and inventory level targets; 2) Iteratively solve the problem using mathematical programming, intelligent optimization, or a hybrid approach to generate multiple candidate production plan solutions; 3) In multi-objective conflict situations, one of the following strategies should be used for solution selection and decision output: • Weighted summation strategy: Combine multiple objective functions linearly according to preset or dynamically adjusted weights, transforming them into a single-objective optimization problem for solution; • Hierarchical optimization strategy: Solve the constraints layer by layer according to the order of primary and secondary objectives, and optimize the secondary objectives while satisfying the higher priority objectives; • Multi-objective screening strategy based on Pareto non-dominated sorting: The candidate solutions are non-dominated, a Pareto front solution set is constructed, and the optimal solution that satisfies the stability index or risk constraint is selected as the output scheme.
[0042] The Pareto non-dominated sorting strategy includes: for any two candidate solutions A and B, if A is not inferior to B in all objective dimensions and is superior to B in at least one objective dimension, then A is determined to dominate B; after filtering out dominated solutions, a set of non-dominated solutions is formed, and the non-dominated solutions are sorted a second time in combination with global stability index or load balance index, and finally the updated production plan is output.
[0043] Furthermore, in step S6, after the data transfer is completed in steps S2 to S5, a global evaluation index system for the production plan is constructed based on the intermediate decision results generated in each step. This system is used to comprehensively evaluate the results of the module's serial operation from three dimensions: algorithm performance, business effectiveness, and scenario stability.
[0044] The overall evaluation index system for the production plan includes: Algorithm performance dimension: The algorithm performance dimension is used to evaluate the rationality and effectiveness of each decision algorithm in steps S2 to S5 in generating intermediate results for the current input data under given parameters and constraints. The algorithm performance dimension includes at least the following indicators: Convergence index: Among them, This represents the evaluation function value of the algorithm in the k-th iteration, where K is the total number of iterations. To prevent extremely small positive numbers with a denominator of 0, k is a variable that takes a value from 1 to K-1. When the convergence index C approaches 1, it indicates that the algorithm has good convergence performance.
[0045] Diversity indicators: Among them, Let represent the feature vector of the plan corresponding to the i-th candidate solution. Let D represent the mean vector of all candidate solutions, and N represent the number of candidate solutions. This diversity index D can help avoid local optima.
[0046] Comprehensive indicators : in, Let j represent the j-th normalized evaluation objective, where j ranges from 1 to M, and M is the number of valid sub-indicators participating in the evaluation under the current algorithm performance dimension. For the corresponding weight coefficients, and .
[0047] After obtaining the above metrics, calculate the comprehensive score for the algorithm's performance dimension: In which P alg This is represented by a score in the algorithm performance dimension. The three factors are weighted coefficients, and their sum is 1. This coefficient is used to evaluate the overall performance of the algorithm in terms of search efficiency, solution space exploration ability, and multi-objective optimization ability.
[0048] Business performance dimension: Used to evaluate output results from the perspective of overall production operation effectiveness, it includes at least the following indicators: Overall cost index C ost : Among them, C prod To represent the cost of steel production and manufacturing, C represents the processing and manufacturing costs incurred by an order at the corresponding production base for completing processes such as steelmaking, continuous casting, rolling, and post-processing. trans C represents the logistics costs incurred during the transfer and transportation of steel between different production bases or between production bases and customers. inv This represents the holding costs incurred when steel is in inventory, including the costs of raw material inventory, work-in-process inventory, finished goods inventory, and surplus inventory.
[0049] Energy consumption indicators : in, Let B represent the unit energy consumption of the p-th process in the b-th base; B represents the total number of production bases participating in the production organization of this planned cycle, b is the base index variable, b=1,2,3,…,B; P represents the total number of processes in the steel production process that participate in the energy consumption calculation, p is the process index variable, p=1,2,3,…,P.
[0050] Delivery time achievement rate (DR): Where, N ontime To complete the order quantity on schedule, N total This represents the total number of orders.
[0051] Capacity utilization and balance indicators: Capacity utilization rate : Capacity balance : in, For actual output, This is the upper limit of production capacity. This represents the average capacity utilization rate.
[0052] After normalizing the above indicators, the business performance score is calculated: In the above formula, cost', Energy', DR', and U b,p ' and Bal' represent the normalized indicators. The weighting coefficients have an algebraic sum of 1. This indicator is used to evaluate the overall effectiveness of production planning in terms of cost control, energy consumption, delivery time assurance, and capacity utilization.
[0053] Scene stability dimension: This evaluation dimension considers the continuous performance of rolling optimization of each module from the perspective of the fluctuation of the output results of each module under different scenarios and loads.
[0054] Indicator volatility : In this formula, j , Indicates the evaluation indicator number, j , =1, 2, 3, ..., M , M , The number of indicators participating in the scenario stability evaluation is represented by t; t represents the number of different operating scenarios or different rolling cycles, t=1, 2, 3, ..., T; T represents the total number of scenarios or rolling cycles selected for statistical analysis. This indicates that in the t-th scene or scrolling cycle, the th... The values of each evaluation indicator; Indicates the first The evaluation metric is the average value across all T scenarios.
[0055] Stability evaluation index Stab: In the formula above, s = 1, 2, 3, ..., N s Indicates the stability sub-index number; N s The number of sub-indicators for stability evaluation; This is a normalized volatility indicator; The corresponding weight coefficients, and satisfying .
[0056] Global comprehensive evaluation function: In a specific example, the global evaluation coordination optimization model constructs a comprehensive evaluation function F: Among them, P alg P is the score for the algorithm's performance dimension.biz The score is for the business performance dimension, the score is for the scenario stability dimension, and w1, w2, and w3 are the weight coefficients.
[0057] After obtaining the calculation result of the comprehensive evaluation function, a parameter correction instruction is generated based on the comprehensive evaluation result, and the parameter correction instruction is applied to the relevant module parameters in steps S2 to S5, thereby forming a feedback adjustment mechanism between modules and realizing global coordination and optimization in the production plan generation process.
[0058] During operation, when the comprehensive evaluation function Below the preset threshold Then, further analysis of the scores of each dimension and the deviation of each evaluation indicator is conducted to determine the module parameters that need to be adjusted.
[0059] Specifically, the deviation of the comprehensive evaluation function is first calculated. : Among them, F target The preset target evaluation value is F, which is the comprehensive evaluation value calculated using formula (13). When When the value is greater than zero, it indicates that the comprehensive evaluation function value is lower than the set target value. At this time, we analyze the changes in the scores of each dimension and further determine the modules that need to be adjusted based on the changes in the indicators of the business performance dimension.
[0060] As shown in formula (13), the comprehensive evaluation function is derived from P alg P biz It consists of three parts: Stab, and the algorithm performance dimension score P. alg This reflects the algorithm's search efficiency and solution quality during the sequential operation of modules. When P alg During the descent, the search efficiency and solution space exploration capabilities of the algorithm can be improved by adjusting the control parameters in the optimization algorithm, such as the candidate solution size, clustering parameters, or number of iterations.
[0061] The scenario stability score (Stab) reflects the stability of the production plan under different operating scenarios. When Stab decreases, production continuity can be improved and production process fluctuations can be reduced by adjusting production batch size, process connection constraints, or production sequencing strategies.
[0062] Business performance score P biz It reflects the effectiveness of the production plan in terms of cost control, energy consumption levels, delivery time achievement rate, equipment utilization rate, and capacity balance. When P bizWhen the performance declines, production decision parameters in steps S2 to S5 are adjusted, such as base allocation weights, surplus material matching thresholds, production line allocation strategies, and batching rules, to improve relevant business indicators and increase the overall evaluation function value. The specific module locations adjusted according to the scoring details are shown below: (1) Feedback adjustment mechanism in step S2: Step S2 is used to make decisions on the allocation of orders among different production bases.
[0063] In this step, the manufacturing costs, transportation costs, process adaptability, and current production load of each production base are comprehensively considered to make base matching decisions for orders and form a base-level order set.
[0064] After step S2 is completed, the global coordination and optimization module calculates a set of evaluation indicators for business performance based on the current production plan, including production cost indicators, energy consumption indicators, delivery date achievement rate indicators, equipment utilization rate indicators, and capacity balance indicators.
[0065] When the system detects significant deviations in relevant indicators of business performance, such as a decline in delivery rate, insufficient capacity utilization, or excessive load differences between production bases, it indicates that the allocation of orders among different production bases may be unreasonable.
[0066] At this point, the global coordination and optimization module generates parameter correction instructions for step S2, adjusting the weight parameters in the base matching decision, such as appropriately increasing the base load penalty weight or decreasing the cost weight, thereby changing the order allocation tendency among different production bases.
[0067] By adjusting the parameters mentioned above, when re-executing the base matching decision, the system will allocate more orders to bases with lower loads and production capacity, thereby improving the load distribution among production bases.
[0068] After the redistribution, the production load of each production base tends to be balanced, which helps to improve the on-time delivery rate, increase equipment utilization, and reduce the capacity balance index, thereby improving the overall business performance score.
[0069] (2) Feedback adjustment mechanism in step S3: Step S3 is used to match order demand with surplus inventory, using surplus inventory as a substitute for order demand to reduce new production demand.
[0070] In this step, the system matches and analyzes the base-level order set with the inventory surplus set. Through the rule knowledge base and matching algorithm, it determines whether the inventory surplus can replace the order demand and generates a surplus material to serve as a solution, thereby forming an order set that includes the net production demand.
[0071] After step S3 is completed, the global coordination and optimization module calculates a set of evaluation indicators for business performance based on the current production plan, including production cost indicators, energy consumption indicators, delivery date achievement rate indicators, equipment utilization rate indicators, and capacity balance indicators.
[0072] When the system detects significant deviations in relevant indicators of business performance, such as high production costs, high energy consumption, or low inventory utilization efficiency, it indicates that the inventory surplus utilization strategy may not be reasonable, and some orders have not made full use of existing inventory resources.
[0073] At this point, the global coordination and optimization module generates parameter correction instructions for step S3, adjusting the relevant parameters in the surplus material matching strategy, such as appropriately relaxing the surplus material matching threshold, increasing the priority of surplus material utilization, or adjusting the weight parameters in the matching rules, thereby expanding the range of surplus materials that can participate in the matching.
[0074] By adjusting the parameters mentioned above, when re-executing the surplus material matching decision, the system can replace more order demands with surplus inventory materials, thereby reducing the amount of new production demand.
[0075] After optimizing the matching of surplus materials, the manufacturing costs and energy consumption levels brought about by new production tasks can be reduced, while improving the utilization efficiency of inventory resources, thereby improving the overall business performance score.
[0076] (3) Feedback adjustment mechanism in step S4: Step S4 is used to complete the allocation of orders among different production lines within each production base.
[0077] In this step, the system constructs an order feature vector based on the product category, specifications, material requirements, delivery constraints, and special process requirements of the order, and forms candidate order clusters through dynamic clustering. Under the premise of meeting the production line process adaptation constraints, the order clusters are configured to the corresponding production lines, thereby forming a production line-level order set.
[0078] After step S4 is completed, the global coordination and optimization module calculates a set of evaluation indicators for business performance based on the production plan results, including production cost indicators, energy consumption indicators, delivery date achievement rate indicators, equipment utilization rate indicators, and capacity balance indicators.
[0079] When the system detects low equipment utilization, large differences in production line load, or high capacity balance index, it indicates that there may be an imbalance in the distribution of orders among different production lines.
[0080] At this point, the global coordination and optimization module generates parameter correction instructions for step S4, adjusting the relevant parameters in the production line allocation strategy, such as adjusting the order clustering threshold, changing the order cluster size, or increasing the production line load balancing weight, thereby changing the configuration of order clusters among different production lines.
[0081] By adjusting the parameters mentioned above, the system can distribute order clusters more evenly across different production lines when re-executing production line allocation decisions, thereby improving the production line load distribution.
[0082] After the redistribution, the load utilization of each production line is more balanced, which helps to improve equipment utilization and reduce the capacity balance index, thereby improving the overall business performance score.
[0083] (4) Feedback adjustment mechanism in step S5: Step S5 is used to complete the collaborative batching of multiple processes such as steelmaking, continuous casting, rolling and post-processing, and form a production batch structure with process continuity through cross-process joint batching.
[0084] In this step, the system combines production line-level order sets into batches based on the process constraints and economic batch requirements between different processes, forming furnace batches, casting batches, rolling batches, and post-processing batches, and establishes batch connection relationships between each process.
[0085] After step S5 is completed, the global coordination and optimization module calculates business performance indicators and scenario stability indicators based on the production plan results.
[0086] When the system detects high energy consumption, large fluctuations in production rhythm, or poor coordination between processes, it indicates that the current batch organization method may be unreasonable.
[0087] At this point, the global coordination and optimization module generates parameter correction instructions for step S5, adjusting the relevant parameters in the cross-process batching strategy, such as adjusting the economic batch size parameter, changing the batch merging rules, or increasing the continuous production constraint weight, thereby changing the way production batches are organized.
[0088] By adjusting the parameters mentioned above, the system can form a more continuous and stable production batch structure when re-executing cross-process batching decisions, thereby reducing equipment changeover frequency and production fluctuations.
[0089] After optimization, energy consumption during the production process can be reduced, and the stability of the production plan under different operating scenarios can be improved, thereby increasing the scenario stability dimension score and further improving the comprehensive evaluation function value.
[0090] As an embodiment of the present invention, the present invention also discloses a global planning optimization system for multi-site, multi-production line, and multi-process steel production, the system being used to implement the method, including: The order and demand management module is used to receive sales order data and perform order standardization processing to obtain an order set. It also provides a structured description of the orders based on steel type, specifications, quantity, delivery date, and special process requirements. Simultaneously, it performs order merging or splitting operations according to production organization needs to form a set of orders to be allocated, and then transmits the set of orders to be allocated to the multi-base order hierarchical processing and allocation module. The multi-base order hierarchical processing and allocation module is used to perform hierarchical clustering and base matching decision on the set of orders to be allocated. During the decision-making process, factors such as the capacity structure, process adaptability range, manufacturing cost, transportation cost, inventory status and key process load level of each production base are comprehensively considered. The optimal production base for an order or order group is determined through the base matching evaluation function, and the base-level order set is output. The inventory surplus material matching module is used to match the base-level order set with the inventory surplus material set in terms of specifications, materials and / or dimensions, generate a surplus material solution, and deduct and mark the net production demand of the base-level order set according to the solution to obtain the base-level order set after surplus material adjustment. The multi-production-line order allocation module is used to perform multi-dimensional feature analysis and dynamic clustering on the order set after the surplus material adjustment within each production base. Under the premise of meeting the adaptation range of production line process, the clustered order clusters are configured to the corresponding production lines according to the order feature vector, production line processing capacity, current load level and energy consumption characteristics, thereby generating a production line-level order set. The multi-process collaborative batching module is used to perform cross-process joint batching of the production line-level order set based on the process constraints and economic batch requirements of steelmaking-continuous casting-rolling-post-processing, generating the batch structure and connection relationship of furnace batch, casting batch, rolling batch or post-processing batch, and obtaining the cross-process batch plan; The global coordination and optimization module is used to calculate global evaluation indicators and output parameter correction instructions for the data transfer process between the multi-base order hierarchical processing and allocation module, the inventory surplus material matching module, the multi-production line order allocation module and the multi-process collaborative batching module, so as to adjust the parameters or constraints of the previous or next step. The plan generation and rolling optimization module is used to generate a production plan scheme that can be optimized and updated for the planning cycle based on the cross-process batch plan and the set of parameters or constraints after global coordination and optimization.
[0091] The following specific examples will be used for illustration: Example: Application of a multi-site, multi-line, multi-process steel production planning system The system in this embodiment includes: an order data input and processing module, a multi-base order allocation module, a surplus material allocation module, a multi-production line allocation module, a multi-process collaborative batching module, a global optimization and adjustment module, and a production plan output and rolling optimization module. Each module performs the following steps: Step 1: Input and processing of order data The system receives and processes data from the sales order system, including customer requirements (steel type, specifications, quantity, etc.) and delivery time requirements. After an order is entered, it first undergoes clustering and splitting by the order hierarchical processing module, merging similar orders to improve production efficiency. Next, the order is transferred to the surplus material matching module, which uses eligible surplus materials as part of the order's production demand, thereby reducing the amount of steel required for new production.
[0092] Step 2: Multi-base order allocation Next, the multi-base order allocation module distributes the processed orders to various production bases. Each base decides whether to accept the order and determines the production sequence based on its current capacity, process adaptability, manufacturing costs, and transportation costs. The order allocation scheme output by this module also takes into account the load balancing issue among the bases and transmits the surplus material usage status to the multi-production line allocation module.
[0093] Step 3: Use leftover materials After multi-site order allocation, the orders are transferred to the surplus material utilization module. The main function of this module is to utilize existing surplus materials in inventory (such as leftover billets and coils from previous production) to replace part of the order demand required for current production. By comparing with the surplus material inventory database, the system automatically determines which orders can partially or fully utilize existing surplus materials.
[0094] Step 4: Multi-production line allocation and load coordination After completing the base allocation, the multi-production-line order allocation module further allocates orders to production lines based on factors such as production line capacity, current load, process requirements, and utilization efficiency of production equipment. The system will generate a production plan for each production base and production line to ensure balanced production line load and maximize capacity utilization.
[0095] Step 5: Multi-process collaborative batching and cross-process scheduling After production line allocation is completed, the multi-process collaborative batching module performs cross-process joint batching of orders on the same production line. This module combines orders on the production line into optimal batches based on the process requirements of different processes (e.g., organizing furnace batches by steel grade and furnace capacity in the steelmaking stage, and sequentially cascading furnaces by casting sequence in the continuous casting stage, etc.), thereby reducing waiting time and mismatch time between processes. This improves production efficiency and reduces losses caused by production line changeovers.
[0096] Step 6: Global Optimization and Feedback Adjustment Throughout the operation of all modules, the global evaluation and coordination optimization model continuously evaluates the output of each stage. The system calculates a comprehensive evaluation result based on multiple dimensions (such as capacity utilization, cost, energy consumption, and delivery time achievement rate) and adjusts the module parameters accordingly. For example, during the order allocation stage, if the capacity load of a certain base is high, the system will allocate more orders to bases with lower loads, thereby optimizing capacity utilization.
[0097] Step 7: Production plan generation and rolling optimization Finally, through multiple iterations and optimizations, the system generates an optimized production plan and feeds it back to the production execution system. This plan will be dynamically adjusted throughout the production cycle to address situations such as order changes, equipment failures, or capacity fluctuations, ensuring high efficiency and stability in production.
[0098] The foregoing description illustrates and describes several preferred embodiments of the present invention. However, as previously stated, it should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the inventive concept described herein through the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.
Claims
1. A global planning optimization method for multi-site, multi-production line, and multi-process steel production, characterized in that, The method includes: S1. Receive sales order data and perform order standardization processing to obtain an order set. Merge or split the order set to form a set of orders to be allocated. S2. Perform hierarchical clustering and base matching decisions on the set of orders to be assigned to obtain a base-level order set and its base affiliation; S3. Match the base-level order set with the inventory surplus material set in terms of specifications, materials and / or dimensions to generate a surplus material solution, and deduct and mark the net production demand of the base-level order set according to this solution to obtain the base-level order set after surplus material adjustment; S4. Perform multidimensional feature analysis and dynamic clustering on the base-level order set after adjusting the surplus materials within each base, and configure the order clusters formed by the clustering results to specific production lines to obtain production line-level order sets; S5. Based on the process constraints and economic batch requirements of steelmaking-continuous casting-rolling-post-processing, the production line-level order set is batched across processes to generate the batch structure and connection relationship of furnace batch, casting batch, rolling batch or post-processing batch, and obtain the cross-process batch plan; S6. Set up a global coordination optimization model, calculate global evaluation indexes for the data transfer process between steps S2 and S5, and output parameter correction instructions to adjust the parameters or constraints of the previous or next step. S7. Based on the cross-process batch plan and the set of parameters or constraints after global coordination and optimization, generate a production plan scheme that can be optimized and updated for the planning cycle.
2. The method according to claim 1, characterized in that, The order set includes steel type, specifications, delivery date, quantity, and special process requirements. The specific steps for merging or splitting the order set are as follows: The order merging includes: clustering and merging orders with the same or compatible steel type, similar specifications, and delivery dates within a preset time window to form candidate production batches; The order splitting includes: for ultra-large batch orders or orders with delivery time spans exceeding a threshold, splitting the order into multiple sub-orders based on base process boundaries, capacity limitations, or delivery time segments, so that each sub-order can be entered into the base allocation decision.
3. The method according to claim 1, characterized in that, The hierarchical clustering and base matching decision is based on the capacity structure, process adaptability range, transportation and manufacturing costs, base inventory and production status of each base. When making decisions, the differences in manufacturing costs of each base, cross-base transportation costs, inventory holding costs, order and base process matching degree, base key process load level and load balance constraints are taken into account to obtain the allocation result structure data that includes the impact of order, base, expected cost and base load.
4. The method according to claim 1, characterized in that, The matching specifically includes: calculating the substitutability ratio of surplus materials to orders, generating surplus materials to act as orders; marking the order being acted as with a result identifier, and passing the mark result identifier along with the order data to step S4.
5. The method according to claim 1, characterized in that, S4 specifically includes: constructing an order feature vector based on the product category, specifications, material requirements, delivery constraints, and special process requirements of the order; under the premise of meeting the production line process adaptation range constraints, dynamically clustering the order feature vector to form a candidate order cluster, and configuring the candidate order cluster to the target production line according to the production line processing capacity, current load level, and energy consumption characteristics, and outputting a production line-level order set.
6. The method according to claim 1, characterized in that, The S5 specifically includes: organizing furnace batches according to steel type and furnace capacity in the steelmaking stage; cascading the furnace batches according to casting sequence in the continuous casting stage; organizing rolling batches according to billet specifications and mill roll changing constraints in the rolling stage; and organizing post-processing batches according to customer specifications and surplus material utilization strategies in the post-processing stage, and generating cross-process connection constraint data that reflects the quantity and sequence matching relationship between processes.
7. The method according to claim 1, characterized in that, S6 specifically includes: first calculating a global evaluation index set that includes at least delivery date achievement rate, capacity utilization rate, equipment utilization rate, capacity balance, energy consumption level and inventory level; then generating parameter correction instructions based on the index set; the parameter correction instructions are used to adjust the clustering threshold, weight coefficient, constraint enable / disable status or assign priority, so that the output of subsequent modules meets the global constraint consistency.
8. The method according to claim 1, characterized in that, The production planning scheme generation includes: calculating the load utilization rate of each process at each base based on cross-process batch planning, identifying the overload risk of bottleneck processes or key equipment; taking capacity balance constraints and equipment utilization rate constraints as global optimization constraints or optimization objectives, and adopting a cross-process hierarchical planning coordination method to transmit constraints and adjust the pace of adjustment step by step, so that the steelmaking pace and the continuous casting times, the continuous casting billet supply and rolling start sequence, and the post-processing capacity meet the executable spatiotemporal matching relationship.
9. The method according to claim 1, characterized in that, The optimization update includes: during the planning period, when order insertion, equipment failure, maintenance plan changes, or surplus material inventory updates cause changes in constraints, dynamically updating the global constraint set and triggering a second solution; the solution adopts mathematical programming, intelligent optimization, or a combination thereof, and when there are multiple conflicting objectives, it adopts weighted summation, hierarchical optimization, or Pareto strategy to output the updated production plan.
10. A global planning and optimization system for multi-site, multi-production line, and multi-process steel production, characterized in that: The system is used to implement the method according to any one of claims 1-9, comprising: The order and demand management module is used to receive sales order data and perform order standardization processing to obtain an order set, and to merge or split the order set to form a set of orders to be allocated. The multi-base order hierarchical processing and allocation module is used to perform hierarchical clustering and base matching decisions on the set of orders to be allocated, so as to obtain a base-level order set and its base affiliation; The inventory surplus material matching module is used to match the base-level order set with the inventory surplus material set in terms of specifications, materials and / or dimensions, generate a surplus material solution, and deduct and mark the net production demand of the base-level order set according to the solution to obtain the base-level order set after surplus material adjustment. The multi-production-line order allocation module is used to perform multi-dimensional feature analysis and dynamic clustering on the base-level order set after adjusting the surplus materials within each base, and to configure the order clusters formed by the clustering results to specific production lines to obtain production-line-level order sets. The multi-process collaborative batching module is used to perform cross-process joint batching of the production line-level order set based on the process constraints and economic batch requirements of steelmaking-continuous casting-rolling-post-processing, generating the batch structure and connection relationship of furnace batch, casting batch, rolling batch or post-processing batch, and obtaining the cross-process batch plan; The global coordination and optimization module is used to calculate global evaluation indicators and output parameter correction instructions for the data transfer process between the multi-base order hierarchical processing and allocation module, the inventory surplus material matching module, the multi-production line order allocation module and the multi-process collaborative batching module, so as to adjust the parameters or constraints of the previous or next step. The plan generation and rolling optimization module is used to generate a production plan scheme that can be optimized and updated for the planning cycle based on the cross-process batch plan and the set of parameters or constraints after global coordination and optimization.
