A decision support system for simulation of logistics scheduling and optimization of production capacity of a metallurgical production line

Abstract

The present application belongs to the technical field of metallurgical production line logistics scheduling, and particularly relates to a kind of metallurgical production line logistics scheduling simulation and capacity optimization decision support system. Including data acquisition and fusion module, logistics scheduling simulation module, capacity optimization decision module, collaborative optimization closed loop module, execution and feedback module, visual simulation and decision support platform module. The present application realizes the deep collaboration of production and logistics, significantly reduces the vehicle waiting time, improves the loading and unloading efficiency and the per capita shipment volume, reduces the logistics cost, and helps the green and low-carbon transformation.

Description

Technical Field

[0001] This invention belongs to the field of logistics scheduling technology for metallurgical production lines, and particularly relates to a simulation and capacity optimization decision support system for logistics scheduling of metallurgical production lines. Background Technology

[0002] Metallurgical enterprises have long production processes, a wide variety of materials, and complex logistics links. Traditional logistics scheduling and management suffer from the following problems:

[0003] 1. Lack of coordination between production planning and transportation needs, resulting in a mismatch between material arrival time and production rhythm, leading to long vehicle waiting times, inventory backlog, or shortages.

[0004] 2. The allocation of resources for vehicle entry and loading / unloading relies heavily on manual experience and lacks scientific decision-making basis, which can easily lead to uneven utilization of loading / unloading equipment and vehicle queues and congestion.

[0005] 3. The production system, warehousing system, and logistics system are independent of each other, and data cannot be shared in real time, making it difficult to achieve global optimization.

[0006] 4. In scenarios where rail and road transport share conveyor belts, frequent process switching prevents the full utilization of loading and unloading capacity; steel coil loading still relies on manual operation, which is inefficient and poses safety hazards.

[0007] 5. The effects of adjusting the scheduling plan cannot be predicted beforehand, and it can only be implemented through trial and error, which is risky and costly.

[0008] Therefore, there is an urgent need for a logistics scheduling optimization system that can achieve deep coupling between production and logistics and has simulation and intelligent decision-making capabilities, so as to improve the overall operating efficiency of metallurgical production lines. Summary of the Invention

[0009] The purpose of this invention is to provide a metallurgical production line logistics scheduling simulation and capacity optimization decision support system to solve the problems existing in the prior art.

[0010] The technical solution adopted by this invention to solve its technical problem is:

[0011] A metallurgical production line logistics scheduling simulation and capacity optimization decision support system includes:

[0012] The data acquisition and fusion module is configured to collect production plan data, material inventory data, in-transit transportation data, and loading and unloading resource data of the metallurgical production line in real time.

[0013] The logistics scheduling simulation module is configured to perform multi-dimensional simulation and deduction of the entire logistics chain of the metallurgical production line based on the data. The simulation and deduction includes at least material shortage prediction simulation, inventory rolling pre-simulation simulation, queuing theory simulation and digital twin simulation.

[0014] The capacity optimization decision module is configured to take the simulation results as input, comprehensively consider transportation costs, transportation resources, inventory constraints, loading and unloading capacity and material priority, solve the optimal resource allocation scheme through operations research optimization algorithm, and output the optimized decision scheme.

[0015] The collaborative optimization closed-loop module is configured to input the optimization decision scheme into the logistics scheduling simulation module for simulation verification, and approximate the global optimal solution through multiple rounds of simulation-optimization iteration;

[0016] The execution and feedback module is configured to send the final optimized scheduling instructions to the execution layer, obtain the execution results through the data acquisition and fusion module, and feed the execution results back to the logistics scheduling simulation module to achieve dynamic correction and adaptive optimization.

[0017] The visualization simulation and decision support platform module is configured to visualize and evaluate the effects of end-to-end logistics scheduling solutions.

[0018] Furthermore, the logistics scheduling simulation module includes:

[0019] The material shortage prediction simulation submodule is configured to predict material shortages within future time windows based on production plans and dynamic BOM tables, and output material demand curves and early warning information.

[0020] The inventory rolling simulation submodule is configured to use the current inventory as a benchmark to continuously predict the inventory change curve of each loading area in the next 24 hours on an hourly basis, simulating the two-way dynamic process of inbound and outbound.

[0021] The queuing theory simulation submodule is configured to establish M / M / c / N / ∞ queuing theory models for different types of materials, dynamically calculate at least service intensity, queue length and waiting time thresholds, and simulate the queuing evolution process.

[0022] The digital twin simulation submodule is configured to build a virtual mirror of the physical world, and uses multimodal perception data to map the status of vehicles, materials, equipment and environment in real time, supporting "what-if" simulation.

[0023] Furthermore, the material shortage prediction simulation submodule is configured as follows:

[0024] Extract relevant data from the existing production system, which includes at least the production planning system, APS scheduling system, raw material storage system and logistics tracking system. The relevant data includes at least the master production schedule, material consumption rate, real-time inventory and in-transit material data.

[0025] The dynamic BOM table is used to calculate the gross demand and demand time of each material. The dynamic BOM table is used to adjust the consumption coefficient in real time according to the process path, specification parameters and historical consumption data.

[0026] By calculating the difference and taking into account the procurement lead time, safety stock threshold, and consumption fluctuation factor, the system dynamically predicts material shortages and their occurrence time, and outputs multi-level early warning information.

[0027] Furthermore, the inventory rolling simulation submodule is configured as follows:

[0028] The rolling plan gathers material specifications, planned rolling volume, and estimated off-line time in real time, on an hourly basis.

[0029] Obtain real-time inventory snapshots from each loading area of ​​the warehouse management system;

[0030] Based on the current time point, the inventory change curves of steel in each loading area and of each specification are projected on an hourly basis for the next 24 hours.

[0031] Access vehicle waybill data and deduct planned shipments from virtual inventory;

[0032] Based on the difference between actual output and planned output, the inventory of pending orders is dynamically adjusted to achieve inventory correction.

[0033] Furthermore, the queuing theory simulation submodule is configured as follows:

[0034] For different types of materials, an M / M / c / N / ∞ queuing model with limited capacity for multiple service stations is established, where c is the number of available unloading ports, N is the maximum waiting vehicle capacity, λ is the vehicle arrival rate, and μ is the service rate.

[0035] Real-time data collection of current queue number of vehicles and historical unloading efficiency in each unloading area; dynamic updates of λ and μ parameters.

[0036] Calculate the service intensity ρ=λ / (c·μ) for each service station, and calculate the average queue length Lq and average waiting time Wq;

[0037] Simulate queue length distribution, waiting time distribution, and system throughput under different scheduling strategies.

[0038] Furthermore, the capacity optimization decision-making module includes:

[0039] The transportation mode and capacity optimization submodule is configured to minimize the overall transportation cost as the objective function. It comprehensively considers the freight rates of road and rail transport, loading and unloading efficiency, transport capacity resources, inventory constraints and material priorities, and solves the optimal transportation plan through operations research optimization algorithms to output the transportation plan.

[0040] The vehicle time-slot reservation and scheduling submodule is configured to identify the available shipment volume for each time slot based on the inventory rolling pre-simulation results, calculate the available reservation capacity by combining the loading and unloading capacity of each time slot, and intelligently allocate vehicle reservation time slots through the reservation allocation algorithm.

[0041] The intelligent loading and unloading scheduling submodule is configured to dynamically calculate the service intensity and waiting time threshold of each unloading area based on queuing theory simulation results. When the unloading area is idle, it automatically triggers the "call a vehicle to enter the plant" command and dynamically schedules vehicles to enter the plant.

[0042] The unmanned loading and unloading operation optimization submodule is configured to establish a vehicle dynamic priority scoring model, realize dynamic allocation of loading positions through a multi-dimensional matching engine, and optimize the scheduling of lifting equipment through a path planning algorithm.

[0043] Furthermore, the vehicle time-slot reservation and dispatch submodule is configured as follows:

[0044] The shippable volume of steel of each specification in each time period is extracted based on the inventory forecast curve;

[0045] Statistics on loading and unloading equipment capacity for each time period are compiled, and the maximum capacity that can be reserved for each time period is calculated.

[0046] A mathematical model for reservation allocation is established with the goal of maximizing overall utility. Considering vehicle priority, reservation time slot preference and capacity constraints, the optimal reservation allocation scheme is solved.

[0047] The remaining available capacity is updated in real time and can be dynamically adjusted.

[0048] Furthermore, the intelligent loading and unloading scheduling submodule is configured as follows:

[0049] Real-time data collection of queue length, equipment status, and operation progress in each unloading area;

[0050] Based on the service intensity ρ and waiting time threshold Wq output by the queuing theory simulation model, the vehicle calling rules of each unloading area are dynamically adjusted.

[0051] When the unloading area is idle, a vehicle call command is triggered, and a vehicle is selected from the corresponding material unloading queue according to priority to enter the plant;

[0052] For scenarios where rail and road transport share the same conveyor belt, the timing of road transport entering the plant is dynamically adjusted to avoid peak unloading times for trains, thereby reducing the number of conveyor belt process switching times.

[0053] Furthermore, the collaborative optimization closed-loop module is configured as follows:

[0054] The optimized decision-making scheme output by the capacity optimization decision-making module is input into the logistics scheduling simulation module;

[0055] In a digital twin simulation environment, a "what-if" simulation is conducted to simulate the vehicle queuing status, loading and unloading area load, inventory level changes, and capacity efficiency indicators after the implementation of the optimized decision-making scheme.

[0056] The simulation evaluation results are fed back to the capacity optimization decision module to adjust and optimize model parameters or constraints.

[0057] Through multiple rounds of simulation and optimization iterations, the global optimal solution is gradually approximated.

[0058] Real-time data collection and deviation analysis of performance data enable automatic adjustment of optimization decision parameters for subsequent time windows, forming an adaptive optimization closed loop.

[0059] Furthermore, the visualization simulation and decision support platform module includes:

[0060] The end-to-end visualization submodule is configured to display vehicle queuing status, loading and unloading area load, dynamic changes in inventory, execution status of dispatch instructions, and capacity efficiency indicators in a three-dimensional visualization format.

[0061] The simulation and deduction submodule is configured to support user-defined scheduling strategy parameters, perform "what-if" simulations, and compare capacity efficiency indicators under different strategies.

[0062] The decision evaluation submodule is configured to evaluate the expected effects and simulation results of the optimized decision scheme from multiple dimensions, and output an evaluation report and optimization suggestions.

[0063] The closed-loop monitoring submodule is configured to monitor the execution layer's job status and deviations in real time, and dynamically display the adaptive optimization process.

[0064] The present invention has the following beneficial effects:

[0065] 1. By using time-slot reservations and dynamic scheduling based on queuing, the dwell time of vehicles entering and leaving the factory has been reduced by more than 30%; by using unmanned loading and route optimization, the average steel shipment volume per person has increased by more than 20%; and by using material shortage prediction and transportation plan optimization, unloading efficiency has increased by more than 10%.

[0066] 2. The multi-objective operations research optimization model comprehensively considers transportation costs, inventory constraints, and carbon emissions, thereby saving warehousing and transportation costs.

[0067] 3. Guided by production planning, transportation planning is driven by material shortage forecasting, and scheduling is guided by inventory pre-planning, forming a two-way dynamic closed loop of "production driving logistics and logistics feeding back into production".

[0068] 4. The simulation module supports "what-if" deduction, which can predict the effect before the solution is implemented and avoid trial and error costs; the optimization decision is based on real-time data and mathematical models, which is superior to traditional experience scheduling.

[0069] 5. Digital twin and unmanned operation technologies reduce human intervention and improve operational safety and stability; the system is replicable and can be extended to other process manufacturing industries.

[0070] 6. By reducing unnecessary vehicle waiting times and optimizing the combination of transportation modes, energy consumption and pollutant emissions can be reduced, thus helping to achieve the "dual carbon" target. Attached Figure Description

[0071] Figure 1 This is the architecture diagram of the metallurgical production line logistics scheduling simulation and capacity optimization decision support system of the present invention. Detailed Implementation

[0072] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0073] like Figure 1 As shown, a metallurgical production line logistics scheduling simulation and capacity optimization decision support system includes a data acquisition and fusion module, a logistics scheduling simulation module, a capacity optimization decision module, a collaborative optimization closed-loop module, an execution and feedback module, and a visualization simulation and decision support platform module.

[0074] I. Taking a steel production base as an example, a data acquisition and fusion module is first built on the industrial internet platform. Through API interfaces, message queues, and database CDC technology, real-time access is provided to data from existing systems, including but not limited to:

[0075] Production planning system: Master production schedule, steel grades, output, planned time, BOM information;

[0076] APS scheduling system: detailed work order information, material requirements, process routes, actual scheduling of each production line (such as blast furnace, converter, rolling line), real-time consumption rate of key materials (such as molten iron, steel billets, various alloys, refractory materials) and work-in-process inventory.

[0077] Raw material warehousing system: Inventory quantity, storage location, batch, warehousing time and status information of various raw and auxiliary materials (such as iron ore, coal, coke, scrap steel, spare parts);

[0078] Logistics tracking system: in-transit material waybill (in-transit status, in-transit volume), estimated arrival time, carrier;

[0079] MES / L2 / L3 system: actual output, consumption rate, and quality assessment;

[0080] IoT sensing devices: factory entrance gate data, weighbridge data, vehicle operation status, LiDAR point cloud, video stream.

[0081] After being cleaned and transformed, all data is stored in the material subject domain data warehouse, providing a unified data foundation for simulation and decision-making.

[0082] II. Logistics Scheduling Simulation Module

[0083] It includes a material shortage prediction simulation submodule, an inventory rolling simulation submodule, a queuing theory simulation submodule, and a digital twin simulation submodule.

[0084] 1. The algorithm flow of the material shortage prediction simulation submodule is as follows:

[0085] (1) Dynamic BOM expansion and gross requirement calculation

[0086] For each future time period k (usually in hours or days), calculate the gross demand G for material i. i,k :

[0087]

[0088] Where: P k Q represents the set of products produced within time period k; p,k The planned output of product p in time period k; Bill of Materials (BOM) i,p,k σ is the dynamic consumption coefficient of product p for material i (which can be corrected in real time based on actual production data, such as adjusting based on the average actual consumption over the past 30 days); i This is the consumption fluctuation factor for material i, used to address demand uncertainty. For example, it is the standard deviation coefficient of historical consumption, which is usually 0.05~0.2.

[0089] (2) Net demand and inventory rolling updates

[0090] Starting from the current time period t0, calculate the inventory level I at the end of each time period sequentially. i,k :I i,k =I i,k-1 +T i,k -G i,k Among them: I i,k-1 The inventory at the start of time period k (for k=t0, I) i,t0-1 (current real-time inventory); T i,k Let T be the expected arrival quantity of material i in time period k (from in-transit orders); if T i,k Delivery can be made in multiple batches within a given timeframe, allowing for further fine-grained processing.

[0091] Simultaneously track available inventory A i,k (i.e., the portion exceeding the safety stock): A i,k =I i,k -SS i , of which SS i Let be the safety stock threshold for material i.

[0092] (3) Gap Time Identification

[0093] The first appearance of I i,k <SS i The time period k is the potential gap time t. s,i If the inventory level remains above the safety stock level within the future window, there will be no gap warning. To ensure timely warnings, a lead time can also be set: when I... i,k Mark the gap time in advance when the price is about to fall below the safety stock level in the later period.

[0094] (4) Analysis of procurement lead time constraints and remedialability

[0095] Record the gap amount: Gap amount = max(0, SS) i -It s,i ), calculate the time from the current time t0 to the gap time t s,i Available time Δt: Δt = t s,i -t0, the lead time for purchasing material i is LT i Compare:

[0096] If Δt>LT i +δ (δ is the buffer time, such as 2 days), then the gap can be avoided through the normal procurement process, triggering a level 1 warning to provide an alert;

[0097] If LT i <Δt≤LT i If +δ is detected, an order must be placed as soon as possible, otherwise it may be too late, triggering a level 2 warning.

[0098] If Δt≤LT i If this happens, normal procurement will be unable to meet the demand, triggering a Level 3 warning and requiring the activation of an emergency plan.

[0099] 2. Inventory Rolling Pre-simulation Submodule

[0100] Using the current time T0 as a baseline, the system performs a 24-hour inventory simulation every hour. It obtains the estimated production time and quantity of each steel specification from the rolling schedule, a current inventory snapshot from the warehousing system, and scheduled shipment plans from the waybill system. The system simulates the inbound (planned production) and outbound (scheduled shipment) processes hourly, generating inventory simulation curves for each loading area. If the actual production of a certain steel specification exceeds the plan, the excess is automatically added to the pending inventory; if there is a shortage, it is deducted from the future pending inventory in chronological order. The simulation results are used to generate recommended available time slots.

[0101] For example, the preview showed that from 14:00 to 16:00 on March 6, there was sufficient stock of hot-rolled coils, and 10 reservation slots could be opened; while from 18:00 to 20:00, the stock was low, and only 3 slots could be opened.

[0102] 3. Queuing Theory Simulation Submodule

[0103] The system establishes an M / M / c / N / ∞ queuing model for each unloading area. Taking the tippler coal unloading area as an example, c = 2 unloading ports, and N = 8 vehicles with a maximum queuing capacity. Real-time data is collected on the vehicle arrival rate λ = 5 vehicles / hour and the average unloading time μ = 0.8 vehicles / hour over the past hour. The calculated service intensity ρ = 5 / (2×0.8) = 3.125 > 1, indicating that the system is overloaded and the queue will continue to grow. Simulations show that the average waiting time will exceed 4 hours. Based on this, the system adjusts the dispatching rules: suspending vehicle dispatch to this area, while coordinating staggered unloading times with rail transport, reducing conveyor belt switching, and improving service efficiency.

[0104] 4. Digital Twin Simulation Submodule

[0105] LiDAR and cameras are deployed in the steel coil storage area to monitor vehicle location, driving status, and steel coil stacking positions in real time. A digital twin is created using a 1:1 scale 3D model, synchronizing physical status every second. When new scheduling strategies need to be evaluated, managers can modify parameters within the digital twin environment (such as adding a vehicle or adjusting parking space allocation rules). The system automatically simulates the operational situation for the next hour, outputting indicators such as congestion probability and equipment utilization rate.

[0106] III. Capacity Optimization Decision Module

[0107] 1. Transportation Mode and Capacity Optimization Submodule

[0108] Using a weekly cycle, a mixed-integer programming model is established based on material shortage forecasts, taking into account port inventory, road and rail transport costs, carbon emission factors, and in-plant loading and unloading capacity. The objective function is to minimize total transportation costs (including freight, loading and unloading fees, and carbon emission costs), with constraints including maximum available transport capacity, inventory level limits, delivery time no later than demand time, and loading and unloading capacity limitations. The mixed-integer programming model is solved using a branch and bound method or a genetic algorithm, and the resulting transport plan is output.

[0109] 2. Vehicle time-slot reservation and dispatch submodule

[0110] Drivers select their expected arrival date via a mobile app. The system then displays the available booking capacity for each time slot based on inventory simulation results. After a driver selects a time slot, the system runs a booking allocation algorithm. Aiming to maximize overall utility, the algorithm solves a mathematical model for booking allocation, constrained by time slot capacity and the set of available vehicle time slots. It comprehensively considers vehicle priority (such as urgent orders and historical fulfillment rates) and time slot capacity, providing confirmation or recommended adjustments. For example, if a driver wants to arrive at 15:00 on March 6th, but that time slot is full, the system recommends 14:00 or 16:00, displaying the available capacity for both time slots. After the driver makes a selection, the system locks the inventory and generates a booking order.

[0111] 3. Intelligent loading and unloading scheduling submodule

[0112] When vehicles arrive within a 5-kilometer radius of the plant, the system dynamically calculates the load of each unloading area based on queuing theory simulation results. When the load ρ in the tipper area drops below 0.7, the system automatically selects the three highest-priority vehicles from the iron ore unloading queue, triggers a "vehicle entry" command, automatically releases them through the gate, and guides them to the designated unloading port via LED screen / APP push notifications. When ρ > 0.9, vehicle dispatch is suspended to avoid excessively long queues. When the waiting time threshold Wq > 2 hours, an early warning is activated, and consideration is given to increasing the service desk or staggered scheduling. Simultaneously, for scenarios where rail and road transport share conveyor belts, the system dynamically adjusts the road transport entry time to avoid peak train unloading periods. If trains arrive in concentrated periods, the system automatically suspends road transport dispatch to avoid frequent conveyor belt switching.

[0113] 4. Optimization Submodule for Unmanned Loading and Unloading Operations

[0114] After the steel coil shipping vehicle enters the unmanned warehouse area, the system generates a loading queue based on the vehicle's waybill (specifications and quantity) and real-time warehouse information using a priority scoring model. This queue is then dynamically allocated to loading positions based on the real-time load of each position. The scoring model parameters include: matching degree (0.4), waiting time (0.3), historical fulfillment rate (0.2), and urgency level (0.1). Once the vehicle arrives at its designated position, the digital twin system activates, automatically planning the route to retrieve the steel coil. The system uses deep learning to identify and verify the steel coil's information code. If the code is correct, the coil is loaded. After loading is complete, the system automatically releases the vehicle and updates the inventory.

[0115] IV. Collaborative Optimization Closed-Loop Module

[0116] The optimized decision scheme output by the capacity optimization decision module is input into the logistics scheduling simulation module. A "what-if" simulation is performed in the digital twin simulation submodule to simulate vehicle queuing status, loading and unloading area load, inventory level changes, and capacity efficiency indicators after the scheme is executed. The simulation evaluation results are compared with the optimization target to calculate the deviations of key indicators, such as average vehicle waiting time, equipment utilization rate, inventory turnover rate, and on-time delivery rate. If the simulation evaluation results do not meet expectations, the results are fed back to the capacity optimization decision module to adjust the optimization model parameters or constraints, such as adjusting the transportation cost weight, modifying the safety stock threshold, and updating the service rate μ of the queuing theory model. Through multiple rounds of simulation-optimization iteration, the global optimal solution is gradually approximated. When the simulation evaluation results meet the preset capacity efficiency threshold, the final optimized decision scheme is output. Every morning, the system performs deviation analysis based on the previous day's execution performance (actual vehicle arrival time, unloading completion time, actual output, etc.). For example, if the actual arrival volume during a certain period is 20% higher than planned, the system automatically increases the λ parameter for the subsequent periods. If the actual service rate of a certain unloading area is lower than expected, μ is decreased, and the queuing theory simulation submodule is rerun to update the scheduling rules. After the optimized decision-making scheme is verified by digital twin, it is updated to the scheduling plan for the day, forming a closed-loop optimization with daily clearing and settlement.

[0117] V. Execution and Feedback Module

[0118] This system is responsible for translating optimized decision-making schemes into executable instructions, such as transportation plans, reservation time slot allocations, vehicle entry scheduling instructions, and unmanned operation instructions. All instructions are transmitted asynchronously and reliably using message queues to ensure no loss or duplication in high-concurrency scenarios. Critical instructions (such as gate access) are equipped with confirmation mechanisms. Upon receiving the instruction, the executing device returns a confirmation signal; if no confirmation is received within the timeout period, a retransmission is triggered.

[0119] It also collects execution data in real time, such as vehicle arrival / departure time, unloading start / end time, unloaded materials / quantities, loading start / end time, loaded materials / quantities, actual output / consumption, equipment status, queue length, etc. The collection frequency varies depending on the data type: key events (such as vehicle arrival) are triggered for real-time upload; status data (such as queue length) is uploaded every 10 seconds; and cumulative data (such as daily output) is synchronized every hour.

[0120] Regularly compare actual performance with the expected values ​​of optimized decision-making plans to identify and quantify deviations. Based on the deviation analysis results, automatically update the model parameters of the logistics scheduling simulation module to achieve dynamic correction.

[0121] After all performance data is timestamped and cleaned, it is stored in a real-time database and synchronized to the material subject domain data warehouse for use by the logistics scheduling simulation module 2 and the collaborative optimization closed-loop module 4.

[0122] VI. Visualization Simulation and Decision Support Platform Module, including:

[0123] 1. End-to-End Visualization Submodule: Displays vehicle queuing status, loading / unloading area load, inventory dynamics, dispatch command execution status, and capacity efficiency indicators in a 3D visualization format on the large screen of the plant dispatch center. Supports multi-view switching, zooming, and clicking to query detailed information, such as vehicle queuing heatmaps, load of each loading / unloading area, inventory level warnings, and dispatch command execution progress.

[0124] 2. Simulation and Deduction Submodule: This module provides a user-friendly interface for the logistics scheduling simulation module. Dispatchers can click on any area to view detailed simulation data and customize scheduling strategy parameters. It allows for "what-if" simulations within a digital twin environment, comparing capacity efficiency indicators under different strategies. For example, dispatchers can adjust parameters such as ride-hailing thresholds (e.g., service intensity ρ upper limit), reservation time slot capacity, production increases / decreases, vehicle delays, and equipment failures. After the dispatcher clicks "Start Simulation," the system uses the current real-world state as the initial state, overlays the user-defined parameter adjustments, and calls various submodules of the logistics scheduling simulation module for rapid simulation. The system instantly displays the simulation results.

[0125] 3. Decision Evaluation Submodule: This module evaluates the expected effects of the optimized decision-making scheme and the simulation results from multiple dimensions, such as efficiency, cost, service level, and energy consumption. It also compares the results with historical data and industry benchmarks, showing the degree of superiority and inferiority, and outputs an evaluation report and optimization suggestions. The report includes key indicator comparisons, risk warnings, and improvement directions.

[0126] Its inputs include optimization decision-making schemes from the capacity optimization decision-making module, simulation results from the logistics scheduling simulation module, and standardized data from the data acquisition and fusion module. These data are then used for evaluation across four dimensions: efficiency, cost, service level, and green / low-carbon performance. A weighted comprehensive scoring method is employed, where standardized indicators for each dimension are summed using a weighted average.

[0127] The weighted comprehensive scoring method is adopted, and the indicators of each dimension are standardized and then summed in a weighted manner:

[0128]

[0129] Where: S is the overall score; D is the set of dimensions (efficiency, cost, service, green); w d Dimension weights (configurable, such as efficiency 0.4, cost 0.3, service 0.2, green 0.1); I d For dimension d, the set of indicators; w d,i x represents the weight of indicator i within the dimension; d,i`norm()` represents the original value of the indicator; `norm()` is the standardization function. By calculating the comprehensive score, multiple alternative solutions can be intuitively mapped to, and compared with historical data and industry benchmarks, thereby assisting decision-makers in selecting the optimal solution.

[0130] Key parameters can be perturbed within a preset range, and the logistics scheduling simulation module is invoked again to quickly calculate the changes in each indicator, output sensitivity curves, and identify which parameters are most sensitive to changes. If a solution is overly sensitive to parameter changes, the system marks it as "high risk." Risks are automatically identified through a rules engine; for example, if material inventory falls below safety stock, a high-level material shortage risk warning is triggered.

[0131] The final output evaluation report includes a comprehensive score for each dimension, which can be displayed as a radar chart; a list of the values ​​of all core indicators, deviations from the target value, and increases or decreases compared to historical values; a list of identified risk items and suggested countermeasures; a sensitivity curve and conclusions; and optimization suggestions (e.g., through back-engineering of predefined rule bases, queuing theory models / shortage models, machine learning, or knowledge bases).

[0132] 4. Closed-Loop Monitoring Submodule: Real-time monitoring of the execution layer's operational status and deviations, such as a vehicle being 15 minutes late or the actual output of hot-rolled coil exceeding the plan by 20 tons. It also dynamically displays the adaptive optimization process, such as showing the iterative process of the collaborative optimization closed-loop module, including the optimization scheme, simulation verification results, and parameter adjustments for each iteration, forming an "iteration history tree" to help users understand how the optimization gradually approaches the optimal point.

[0133] The above content is displayed on the large screen in the factory's dispatch center. Dispatchers can monitor the status of the entire logistics chain in real time and make decisions based on system suggestions.

[0134] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the concept and scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the design concept of the present invention should fall within the protection scope of the present invention.

[0135] The technologies, shapes, and structures not described in detail in this invention are all known technologies.

Claims

1. A simulation and capacity optimization decision support system for logistics scheduling in a metallurgical production line, characterized in that, include: The data acquisition and fusion module is configured to collect production plan data, material inventory data, in-transit transportation data, and loading and unloading resource data of the metallurgical production line in real time. The logistics scheduling simulation module is configured to perform multi-dimensional simulation and deduction of the entire logistics chain of the metallurgical production line based on the data. The simulation and deduction includes at least material shortage prediction simulation, inventory rolling pre-simulation simulation, queuing theory simulation and digital twin simulation. The capacity optimization decision module is configured to take the simulation results as input, comprehensively consider transportation costs, transportation resources, inventory constraints, loading and unloading capacity and material priority, solve the optimal resource allocation scheme through operations research optimization algorithm, and output the optimized decision scheme. The collaborative optimization closed-loop module is configured to input the optimization decision scheme into the logistics scheduling simulation module for simulation verification, and approximate the global optimal solution through multiple rounds of simulation-optimization iteration; The execution and feedback module is configured to send the final optimized scheduling instructions to the execution layer, obtain the execution results through the data acquisition and fusion module, and feed the execution results back to the logistics scheduling simulation module to achieve dynamic correction and adaptive optimization. The visualization simulation and decision support platform module is configured to visualize and evaluate the effects of end-to-end logistics scheduling solutions.

2. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 1, characterized in that, The logistics scheduling simulation module includes: The material shortage prediction simulation submodule is configured to predict material shortages within future time windows based on production plans and dynamic BOM tables, and output material demand curves and early warning information. The inventory rolling simulation submodule is configured to use the current inventory as a benchmark to continuously predict the inventory change curve of each loading area in the next 24 hours on an hourly basis, simulating the two-way dynamic process of inbound and outbound. The queuing theory simulation submodule is configured to establish M / M / c / N / ∞ queuing theory models for different types of materials, dynamically calculate at least service intensity, queue length and waiting time thresholds, and simulate the queuing evolution process. The digital twin simulation submodule is configured to build a virtual mirror of the physical world, and uses multimodal perception data to map the status of vehicles, materials, equipment and environment in real time, supporting "what-if" simulation.

3. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 2, characterized in that, The material shortage prediction simulation submodule is configured as follows: Extract relevant data from the existing production system, which includes at least the production planning system, APS scheduling system, raw material storage system and logistics tracking system. The relevant data includes at least the master production schedule, material consumption rate, real-time inventory and in-transit material data. The gross demand and demand time of each material are calculated based on the dynamic BOM table. The dynamic BOM table is used to adjust the consumption coefficient in real time according to the process path, specification parameters and historical consumption data. By calculating the difference and taking into account the procurement lead time, safety stock threshold, and consumption fluctuation factor, the system dynamically predicts material shortages and their occurrence time, and outputs multi-level early warning information.

4. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 2, characterized in that, The inventory rolling pre-simulation submodule is configured as follows: The rolling plan gathers material specifications, planned rolling volume, and estimated off-line time in real time, on an hourly basis. Obtain real-time inventory snapshots from each loading area of ​​the warehouse management system; Based on the current time point, the inventory change curves of steel in each loading area and of each specification are projected on an hourly basis for the next 24 hours. Access vehicle waybill data for sales and deduct planned shipments from virtual inventory; Based on the difference between actual and planned production, the inventory of pending orders is dynamically adjusted to achieve inventory correction.

5. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 2, characterized in that, The queuing theory simulation submodule is configured as follows: For different types of materials, an M / M / c / N / ∞ queuing model with limited capacity for multiple service stations is established, where c is the number of available unloading ports, N is the maximum waiting vehicle capacity, λ is the vehicle arrival rate, and μ is the service rate. Real-time data collection of current queue number of vehicles and historical unloading efficiency in each unloading area; dynamic updates of λ and μ parameters. Calculate the service intensity ρ=λ / (c·μ) for each service station, and calculate the average queue length Lq and average waiting time Wq; Simulate queue length distribution, waiting time distribution, and system throughput under different scheduling strategies.

6. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 1, characterized in that, The capacity optimization decision module includes: The transportation mode and capacity optimization submodule is configured to minimize the overall transportation cost as the objective function. It comprehensively considers the freight rates of road and rail transport, loading and unloading efficiency, transport capacity resources, inventory constraints and material priorities, and solves the optimal transportation plan through operations research optimization algorithms to output the transportation plan. The vehicle time-slot reservation and scheduling submodule is configured to identify the available shipment volume for each time slot based on the inventory rolling pre-simulation results, calculate the available reservation capacity by combining the loading and unloading capacity of each time slot, and intelligently allocate vehicle reservation time slots through the reservation allocation algorithm. The intelligent loading and unloading scheduling submodule is configured to dynamically calculate the service intensity and waiting time threshold of each unloading area based on queuing theory simulation results. When the unloading area is idle, it automatically triggers the "call a vehicle to enter the plant" command and dynamically schedules vehicles to enter the plant. The unmanned loading and unloading operation optimization submodule is configured to establish a vehicle dynamic priority scoring model, realize dynamic allocation of loading positions through a multi-dimensional matching engine, and optimize the scheduling of lifting equipment through a path planning algorithm.

7. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 6, characterized in that, The vehicle time-slot reservation and scheduling submodule is configured as follows: The shippable volume of steel of each specification in each time period is extracted based on the inventory forecast curve; Statistics on loading and unloading equipment capacity for each time period are compiled, and the maximum capacity that can be reserved for each time period is calculated. A mathematical model for reservation allocation is established with the goal of maximizing overall utility. Considering vehicle priority, reservation time slot preference and capacity constraints, the optimal reservation allocation scheme is solved. The remaining available capacity is updated in real time and can be dynamically adjusted.

8. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 6, characterized in that, The intelligent loading and unloading scheduling submodule is configured as follows: Real-time data collection of queue length, equipment status, and operation progress in each unloading area; Based on the service intensity ρ and waiting time threshold Wq output by the queuing theory simulation model, the vehicle calling rules of each unloading area are dynamically adjusted. When the unloading area is idle, a vehicle call command is triggered, and a vehicle is selected from the corresponding material unloading queue according to priority to enter the plant; For scenarios where rail and road transport share the same conveyor belt, the timing of road transport entering the plant is dynamically adjusted to avoid peak unloading times for trains, thereby reducing the number of conveyor belt process switching times.

9. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 1, characterized in that, The collaborative optimization closed-loop module is configured as follows: The optimized decision-making scheme output by the capacity optimization decision-making module is input into the logistics scheduling simulation module; In a digital twin simulation environment, "what-if" simulations are conducted to simulate the vehicle queuing status, loading and unloading area load, inventory level changes, and capacity efficiency indicators after the implementation of optimized decision-making schemes. The simulation evaluation results are fed back to the capacity optimization decision module to adjust and optimize model parameters or constraints. Through multiple rounds of simulation and optimization iterations, the global optimal solution is gradually approximated. Real-time data collection and deviation analysis of performance data enable automatic adjustment of optimization decision parameters for subsequent time windows, forming an adaptive optimization closed loop.

10. The metallurgical production line logistics scheduling simulation and capacity optimization decision support system according to claim 1, characterized in that, The visualization simulation and decision support platform module includes: The end-to-end visualization submodule is configured to display vehicle queuing status, loading and unloading area load, dynamic changes in inventory, execution status of dispatch instructions, and capacity efficiency indicators in a three-dimensional visualization format. The simulation and deduction submodule is configured to support user-defined scheduling strategy parameters, perform "what-if" simulations, and compare capacity efficiency indicators under different strategies. The decision evaluation submodule is configured to evaluate the expected effects and simulation results of the optimized decision scheme from multiple dimensions, and output an evaluation report and optimization suggestions. The closed-loop monitoring submodule is configured to monitor the execution layer's job status and deviations in real time, and dynamically display the adaptive optimization process.

Images