Dynamic scheduling system and method for intelligent warehouse of waste home appliance raw materials based on digital twinning

Abstract

This invention relates to the field of warehouse scheduling and management technology, specifically to a dynamic scheduling system and method for intelligent warehousing of waste household appliance raw materials based on digital twins. The invention constructs a three-dimensional digital twin model of all elements of the waste household appliance raw material warehouse, achieving real-time bidirectional data synchronization between the physical warehouse and the virtual model; establishes a multi-dimensional outbound priority evaluation system including category weight, quantity, dismantling urgency, and storage duration; employs an improved genetic algorithm incorporating adaptive crossover and mutation probabilities to solve a multi-objective scheduling optimization problem that considers scheduling time, travel distance, and energy consumption; and establishes a graded response mechanism for abnormal events and a dynamic rolling time-domain rescheduling mechanism, combining multi-source heterogeneous data fusion and virtual simulation verification to achieve closed-loop scheduling control. This invention achieves integrated collaboration between storage location management and outbound scheduling, improving warehouse scheduling efficiency and resource allocation levels, ensuring timely disposal of raw materials with high environmental risks, and enhancing the system's dynamic response capability and operational stability.

Description

Technical Field

[0001] This invention relates to the field of warehouse scheduling and management technology, specifically to a dynamic scheduling system and method for intelligent warehousing of waste household appliance raw materials based on digital twins. Background Technology

[0002] With the rapid development of my country's waste household appliance (WHO) recycling and processing industry, the warehousing of WHO raw materials, as a core link in the recycling industry chain, directly affects the capacity and environmental compliance of the entire dismantling production line due to its scheduling efficiency. WHO raw materials are characterized by their diverse categories, large differences in volume and weight, varying environmental risk levels, and dynamic fluctuations in dismantling demand. Traditional manual scheduling methods suffer from low efficiency, slow response, and unreasonable resource allocation, and can no longer meet the needs of large-scale, intelligent WHO recycling and processing.

[0003] The integration of digital twin and intelligent algorithm technologies provides a new solution for warehouse scheduling optimization. Patent application CN121810178A discloses an intelligent warehouse dynamic storage location optimization system and method based on AI and digital twins. This technology acquires shelf inventory status information, establishes constraint functions for the nearest goods storage distance, shelf load-bearing capacity, and shelf stability, constructs a multi-objective optimization objective function, and uses a genetic algorithm to solve the particle population to obtain the optimal storage location allocation scheme. The storage location model of the goods to be stored is highlighted in the digital twin model, which can improve inbound and outbound efficiency while ensuring the safety of goods storage.

[0004] However, the aforementioned existing technologies still have the following shortcomings: First, the technology only focuses on the optimization of storage location allocation during the goods receiving stage, without addressing the dynamic scheduling problem during the warehouse outbound stage, and cannot adapt to complex scenarios such as dynamic changes in dismantling orders and the insertion of emergency environmental disposal tasks in the storage of waste household appliance raw materials; Second, the genetic algorithm it uses uses fixed crossover and mutation probabilities, which is prone to premature convergence and insufficient global search capabilities, making it difficult to quickly solve multi-objective and multi-constraint warehouse scheduling optimization problems; Third, it lacks a sound abnormal event hierarchical response and real-time rescheduling mechanism, and when abnormal situations such as equipment failure or raw material damage occur, it cannot quickly adjust the scheduling plan, which can easily lead to the interruption of warehousing operations; Fourth, it does not establish a multi-dimensional outbound priority evaluation system based on the characteristics of waste household appliance raw materials, and cannot comprehensively consider core factors such as the reuse value of raw materials, environmental risk level, and dismantling urgency, which can easily lead to problems such as storage timeouts of highly polluting raw materials and delays in core dismantling orders.

[0005] Therefore, there is an urgent need to develop a dynamic scheduling system and method specifically designed for the storage characteristics of waste household appliance raw materials, so as to achieve integrated coordination of storage location management and outbound scheduling, and improve the dynamic response capability and resource allocation efficiency of the storage system. Summary of the Invention

[0006] The purpose of this invention is to provide a dynamic scheduling system and method for intelligent warehousing of waste household appliance raw materials based on digital twins, so as to solve the problems mentioned in the background art that the existing technology only focuses on optimizing the allocation of storage locations upon entry, lacks dynamic scheduling upon exit, has insufficient abnormal response capability, and is not adapted to the characteristics of waste household appliance raw materials.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins, characterized by comprising: The digital twin warehouse modeling module is used to construct a three-dimensional digital twin model of waste household appliance raw material warehouse. The three-dimensional digital twin model includes a warehouse physical space mapping unit, a raw material attribute mapping unit, a warehouse equipment mapping unit, and a virtual simulation verification unit. The multi-source heterogeneous data sensing module is used to collect raw material status data, equipment operation data and environmental data in the warehouse in real time, and synchronize the collected data to the three-dimensional digital twin model at a preset frequency; The dynamic priority calculation module is used to calculate the outbound priority of each batch of raw materials based on the type, quantity, urgency of dismantling, and storage duration of the raw materials using a multi-dimensional priority calculation formula. The hybrid intelligent scheduling module is used to solve the multi-objective optimization problem of warehouse scheduling based on real-time status data of the three-dimensional digital twin model and the outbound priority of each batch of raw materials, and to generate an initial scheduling scheme by using an improved adaptive genetic algorithm. The real-time rescheduling trigger module is used to monitor various abnormal events in the warehouse. When an abnormal event occurs, it triggers the corresponding level of real-time rescheduling process according to the severity of the abnormal event. The scheduling execution control module is used to convert scheduling schemes into standardized control commands, which are then sent to execution equipment such as AGVs, stacker cranes, conveyors, and palletizing robots in the warehouse, and the execution status is fed back to the three-dimensional digital twin model in real time.

[0008] Preferably, the multi-dimensional priority calculation formula in the dynamic priority calculation module is as follows: ; in, For the first The outbound priority of each raw material batch, with a value range of [value range missing]. ; For the first The category weighting coefficient for each batch of raw materials is determined based on the metal content, environmental risk level, and reuse value of the raw materials, with a range of values ​​ranging from [value missing]. ; For the first The normalized quantity value of each raw material batch is calculated using the following formula: , This represents the maximum storage capacity per batch. For the first The urgency coefficient for dismantling each batch of raw materials is determined based on the real-time capacity gap of the dismantling line and the order delivery deadline, with a value range of [value missing]. ; For the first The normalized value of the storage duration for each batch of raw materials is calculated using the following formula: , This refers to the maximum permissible storage time for raw materials; Let be the weight coefficient, and satisfy... .

[0009] Preferably, the multi-objective fitness function in the improved adaptive genetic algorithm used by the hybrid intelligent scheduling module is: ; in, This is the fitness value; The total completion time for all scheduled tasks; The total travel distance of all executing devices; The total energy consumption of all executing devices is calculated using the following formula: , The total number of devices to be executed. For the first The load capacity of the equipment For the first The travel distance of the equipment. Energy consumption coefficient per unit load per unit distance The standby power consumption coefficient per unit time. For the first Standby time of the device; Let be the target weight coefficient, and satisfy... .

[0010] Preferably, the improved adaptive genetic algorithm introduces both adaptive crossover probability and adaptive mutation probability, wherein the adaptive mutation probability is calculated using the following formula: ; in, For adaptive mutation probability; The maximum mutation probability is given by a value of [value]. ; To minimize the mutation probability, the value is [value]. ; This represents the average fitness value of the current population. The fitness value of the individual to be mutated; This represents the minimum fitness value of the current population.

[0011] Preferably, the real-time rescheduling triggering module adopts a dynamic scrolling time-domain optimization strategy, setting the basic scrolling window length to [value missing]. When an abnormal event occurs, it is first classified into three levels—minor, moderate, and severe—according to a preset abnormality classification standard. Then, the length of the scrolling window is dynamically adjusted: for minor abnormalities... When moderate abnormality When there is a serious abnormality Simultaneously, a task priority filtering mechanism is introduced, only processing tasks with an outbound priority higher than a preset threshold. The tasks are re-optimized, while tasks outside the rolling window and with a priority below the threshold retain their original scheduling scheme.

[0012] Preferably, the multi-source heterogeneous data sensing module includes an RFID tag reader, a 3D vision sensor, a weight sensor, a temperature and humidity sensor, and a smoke sensor. The multi-source heterogeneous data sensing module uses DS evidence theory to fuse the raw material data collected by the RFID tag reader, 3D vision sensor, and weight sensor to improve the raw material identification accuracy. The data synchronization frequency is set to: equipment operating status data and raw material location data every [time / period]. Synchronize once, environmental data every Once synchronized, raw material attribute data is synchronized in real time during both inbound and outbound processes.

[0013] On the other hand, the present invention also provides a dynamic scheduling method for intelligent warehousing of waste household appliance raw materials based on digital twins. This method is implemented based on the above-mentioned dynamic scheduling system and includes the following steps: S1: Construct a three-dimensional digital twin model of the waste household appliance raw material storage and establish a real-time data mapping channel between the physical storage and the digital twin model; S2: Real-time collection of raw material status data, equipment operation data, and environmental data within the warehouse via a multi-source heterogeneous data sensing module, and updating of these data to the three-dimensional digital twin model; S3: Calculate the outbound priority of each batch of raw materials based on the type, quantity, urgency of dismantling, and storage duration using a multi-dimensional priority calculation formula; S4: Based on the real-time status data of the three-dimensional digital twin model and the outbound priority of each batch of raw materials, an improved adaptive genetic algorithm is used to solve the multi-objective optimization problem of warehouse scheduling and generate an initial scheduling scheme. S5: Monitors various abnormal events in the warehouse. When an abnormal event occurs, it triggers the corresponding level of real-time rescheduling process according to the severity of the abnormal event and generates an updated scheduling plan. S6: Convert the scheduling scheme into standardized control commands, issue them to the execution equipment in the warehouse, and provide real-time feedback on the execution status to the three-dimensional digital twin model.

[0014] Preferably, step S3 specifically includes the following steps: S31: Obtain the category information, actual quantity, dismantling line capacity gap data, and stored duration data of all raw material batches to be shipped from the three-dimensional digital twin model; S32: Determine the category weight coefficient for each batch of raw materials based on the preset category weight table. The urgency coefficient for dismantling each batch of raw materials is determined based on the dismantling line capacity gap and order delivery deadlines. ; S33: Calculate the normalized quantity values ​​for each batch of raw materials. and storage duration normalized value ; S34: Substitute into the multi-dimensional priority calculation formula to calculate the outbound priority of each raw material batch. The raw material batches are sorted from highest to lowest priority.

[0015] Preferably, step S4 specifically includes the following steps: S41: Encode the scheduling task and execution device using integer encoding to generate an initial population, with the population size set to [value missing]. ; S42: Calculate the fitness value of each individual in the population based on the multi-objective fitness function; S43: Use a roulette wheel selection method combined with an elite retention strategy to select the option with the highest fitness value. Individuals directly enter the next generation; S44: Perform crossover and mutation operations on individuals in the population based on adaptive crossover and adaptive mutation probabilities; S45: Determine if the preset number of iterations has been reached. Or fitness value convergence threshold If so, output the optimal scheduling scheme; otherwise, return to step S42 to continue iterating.

[0016] Preferably, step S5 specifically includes the following steps: S51: Real-time monitoring of abnormal events within the warehouse; when an abnormal event is detected, it is classified and its level is determined. S52: Update the corresponding status data of the 3D digital twin model according to the type and level of the abnormal event, including equipment failure status, raw material damage status and task change status; S53: Determine the length of the rolling window and the scope of tasks participating in rescheduling based on the level of the abnormal event, and filter out tasks with priority below the threshold; S54: Call the improved adaptive genetic algorithm to re-optimize and solve the tasks within the rolling window, and generate an updated scheduling scheme; S55: The updated scheduling scheme is sent to the scheduling execution control module, and the virtual scheduling status in the three-dimensional digital twin model is updated at the same time.

[0017] Compared with the prior art, the beneficial effects of the present invention are: (1) This invention constructs a three-dimensional digital twin model of all elements of waste household appliance raw material storage and combines it with a multi-dimensional dynamic outbound priority calculation method to achieve integrated collaborative optimization of storage location management and outbound scheduling. The digital twin model realizes real-time bidirectional mapping of physical storage space, raw material attributes, and equipment status, providing full-dimensional and high-precision status data support for scheduling decisions. In view of the characteristics of the complex categories and large differences in environmental risk levels of waste household appliance raw materials, a multi-dimensional priority evaluation system including category weight, quantity, dismantling urgency, and storage time has been established. This system can take into account the reuse value of raw materials, environmental disposal requirements, and dismantling production needs, avoiding problems such as storage timeouts of high-pollution raw materials and delays in core dismantling orders. It also makes up for the shortcomings of existing technologies that only focus on the allocation of storage locations and lack outbound scheduling optimization.

[0018] (2) This invention employs an improved adaptive genetic algorithm to solve the multi-objective optimization problem of warehouse scheduling. By introducing adaptive crossover probability and adaptive mutation probability, the algorithm parameters are dynamically adjusted according to the real-time fitness of the population, effectively solving the problems of premature convergence and low search efficiency in the later stages of traditional genetic algorithms, thus improving the solution accuracy and convergence speed of multi-objective optimization problems. The algorithm constructs a multi-objective fitness function that includes total scheduling completion time, total equipment travel distance, and total equipment energy consumption. It can reduce energy consumption and equipment wear in warehouse operations while ensuring scheduling efficiency, achieving multi-objective optimal allocation of warehouse resources and overcoming the limitations of existing technologies that use fixed-parameter genetic algorithms for solving problems.

[0019] (3) This invention establishes a graded response and dynamic rolling time-domain rescheduling mechanism for abnormal events. Warehouse abnormal events are divided into three levels: minor, moderate, and severe. The length of the rolling window is dynamically adjusted for different levels of abnormality, and a task priority filtering mechanism is used to select the scope of rescheduling tasks. This mechanism can respond quickly when abnormal events occur, adjust the scheduling plan in a timely manner to ensure the continuity of warehousing operations, and effectively reduce the computational load of rescheduling by limiting the scope of rescheduling tasks, thereby improving the scheduling response speed. This solves the problem that existing technologies lack a complete abnormality handling mechanism and are prone to operation interruptions when abnormalities occur.

[0020] (4) This invention achieves closed-loop control of the warehouse scheduling process through multi-source heterogeneous data fusion technology and virtual simulation verification. The multi-source heterogeneous data sensing module uses DS evidence theory to fuse data from multiple sensors, improving the accuracy of raw material identification and status monitoring. Before the scheduling scheme is executed, virtual simulation verification is performed in the digital twin model, which can detect problems such as equipment collisions, path conflicts, and operation deadlocks in advance, avoiding ineffective scheduling and operation accidents. During the scheduling execution, the equipment operating status and task progress are fed back in real time, realizing dynamic synchronization between physical operations and virtual models, improving the reliability and stability of the warehouse scheduling system, and meeting the actual needs of large-scale and intelligent operation of waste household appliance raw material storage. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are explained in detail together with the embodiments of the invention, but do not constitute a limitation thereof.

[0022] Figure 1 This is a module architecture diagram of the intelligent warehousing and dynamic scheduling system for waste household appliance raw materials based on digital twins, as described in this invention. Figure 2 This is a flowchart of the intelligent warehousing and dynamic scheduling method for waste household appliance raw materials based on digital twins, as described in this invention. Figure 3 This is a flowchart of the digital twin modeling and data perception process of the present invention; Figure 4 This is a flowchart of the dynamic priority calculation process of the present invention; Figure 5 This is a flowchart of the hybrid intelligent scheduling and genetic algorithm of the present invention; Figure 6 This is a flowchart of the real-time rescheduling triggering and execution process of the present invention. Detailed Implementation

[0023] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0024] like Figures 1-6As shown, this invention's dynamic scheduling system for intelligent warehousing of waste household appliance raw materials, based on digital twins, achieves real-time data interaction and state synchronization between the physical entity and the virtual model by constructing a full-element three-dimensional digital twin model of the physical warehouse. The system employs multi-source heterogeneous data sensing technology to collect full-dimensional operational data of the warehouse, determines the raw material outbound order through multi-dimensional dynamic priority calculation, solves the multi-objective scheduling optimization problem using an improved adaptive genetic algorithm, and triggers hierarchical real-time rescheduling based on the level of abnormal events, ultimately achieving efficient, intelligent, and dynamic scheduling of waste household appliance raw material warehousing.

[0025] The digital twin warehouse modeling module uses 3D laser scanning technology combined with BIM modeling methods to construct a high-precision 3D digital twin model of waste household appliance raw material warehouses. The model includes warehouse physical space mapping units, raw material attribute mapping units, warehouse equipment mapping units, and virtual simulation verification units.

[0026] The warehouse physical space mapping unit accurately maps physical space information within the warehouse, such as shelf layout, aisle dimensions, entrance and exit locations, and work area divisions. The raw material attribute mapping unit stores and updates in real-time the attribute information of all raw material batches, including category, quantity, arrival time, storage location, metal content, and environmental risk level. The warehouse equipment mapping unit maps the geometric parameters, kinematic parameters, and operating status parameters of execution equipment such as AGVs, stacker cranes, conveyors, and palletizing robots. Before the scheduling plan is executed, the virtual simulation verification unit simulates the execution process of the scheduling plan in a 3D digital twin model, detecting issues such as equipment collisions, path conflicts, and job deadlocks. If the simulation verification passes, the scheduling plan is executed; otherwise, a new scheduling plan is generated.

[0027] Multi-source heterogeneous data sensing modules are deployed at various key nodes in the warehouse, including RFID tag readers, 3D vision sensors, weight sensors, temperature and humidity sensors, and smoke sensors. RFID tag readers are deployed at the inbound and outbound entrances and shelf nodes to read the information from electronic tags affixed to waste household appliance raw materials. 3D vision sensors are deployed at aisle entrances and above work areas to identify the appearance characteristics and real-time location of raw materials. Weight sensors are deployed at inbound and outbound stations to detect the actual weight of raw material batches. Temperature and humidity sensors and smoke sensors are evenly distributed throughout the warehouse to monitor environmental temperature and humidity and fire hazards.

[0028] The multi-source heterogeneous data sensing module uses DS evidence theory to fuse raw material data collected by RFID tag readers, 3D vision sensors, and weight sensors. The data synchronization frequency is set to synchronize equipment operating status data and raw material location data every 100ms, environmental data every 1min, and raw material attribute data in real time during warehousing and outbound processes.

[0029] The dynamic priority calculation module calculates the outbound priority of each raw material batch based on its category, quantity, urgency of dismantling, and storage duration using a multi-dimensional priority calculation formula. The calculation formula is as follows: ; in, For the first The outbound priority of each raw material batch, with a value range of [value range missing]. . For the first The category weighting coefficient for each batch of raw materials is determined based on the metal content, environmental risk level, and reuse value of the raw materials, with a range of values ​​ranging from [value missing]. . For the first The normalized quantity value of each raw material batch is calculated using the following formula: , This represents the maximum storage capacity for a single batch. For the first The urgency coefficient for dismantling each batch of raw materials is determined based on the real-time capacity gap of the dismantling line and the order delivery deadline, with a value range of [value missing]. . For the first The normalized value of the storage duration for each batch of raw materials is calculated using the following formula: , This is the maximum permissible storage time for raw materials. Let be the weight coefficient, and satisfy... In this embodiment, , , , .

[0030] The hybrid intelligent scheduling module, based on real-time status data from a 3D digital twin model and the outbound priority of each raw material batch, employs an improved adaptive genetic algorithm to solve the multi-objective optimization problem of warehouse scheduling. The algorithm's multi-objective fitness function is: ; in, This is the fitness value. This represents the total completion time for all scheduled tasks. This represents the total travel distance of all executing devices. The total energy consumption of all executing devices is calculated using the following formula: , The total number of devices to be executed. For the first The load capacity of the equipment For the first The travel distance of the equipment. Energy consumption coefficient per unit load per unit distance The standby power consumption coefficient per unit time. For the first Standby time of the device. Let be the target weight coefficient, and satisfy... .

[0031] The improved adaptive genetic algorithm introduces both adaptive crossover probability and adaptive mutation probability, effectively addressing the shortcomings of traditional genetic algorithms, such as premature convergence and low search efficiency in the later stages. The formula for calculating the adaptive crossover probability is: ; in, For adaptive crossover probability, For the maximum crossover probability, Minimum crossover probability; For adaptive mutation probability, The maximum mutation probability, The minimum mutation probability; This represents the average fitness value of the current population. The larger fitness value among the two crossover individuals. The fitness value of the individual to be mutated. This represents the minimum fitness value of the current population.

[0032] This algorithm uses integer encoding to encode the scheduling task and the execution device. The preset population size is 100, the maximum number of iterations is 200, and the fitness convergence threshold is set to... The selection operation employs a roulette wheel selection method combined with an elite retention strategy, preserving the top 10% of individuals with the highest fitness values ​​in the population directly into the next generation, maximizing the retention of high-quality scheduling solutions while balancing the algorithm's convergence speed and global search capability.

[0033] The real-time rescheduling trigger module monitors abnormal events throughout the entire warehousing operation process around the clock. It can accurately identify various operational disturbances and handle them in a tiered manner, ensuring the dynamic stability of warehousing scheduling. The system classifies warehousing abnormal events into three levels: minor, moderate, and severe, with different rescheduling strategies corresponding to different levels of abnormalities. Minor abnormalities include minor equipment stalls, raw material position deviations, and slight exceedances of temperature and humidity in the warehousing environment, which do not affect core operations. Moderate abnormalities include single handling equipment failures, minor raw material damage, and temporary changes to routine outbound orders, which are considered moderate disturbances. Severe abnormalities include multiple equipment shutdowns, large-scale raw material damage, insertion of emergency environmental disposal orders, and early warnings of warehousing safety hazards, which are considered major disturbances.

[0034] This module employs a dynamic scrolling time-domain optimization strategy, setting the base scrolling window length. The window length can be dynamically adjusted based on the anomaly level to accurately pinpoint the scope of rescheduled tasks. Under minor anomaly conditions, the scrolling window length is set to... Optimizations are only applied to a small number of currently available tasks; under moderately abnormal operating conditions, the scroll window length remains at the base value. It adapts to the needs of routine operation disturbance optimization; under severe abnormal operating conditions, the scrolling window length is expanded to This covers more pending tasks and ensures effective scheduling in emergency scenarios. It also allows setting task priority filtering thresholds. Only core tasks with outbound priority above the threshold are rescheduled and optimized, while regular standby tasks with priority below the threshold retain their original scheduling scheme, effectively reducing the computational load of rescheduling and improving scheduling response speed.

[0035] The scheduling and execution control module, as the output terminal of the system scheduling instructions, is responsible for converting the optimal scheduling scheme generated by the virtual model into standardized equipment control instructions, thereby realizing the automated collaborative operation of warehousing equipment. The standardized control instructions include core parameters such as equipment number, task type, raw material storage and retrieval location, operation sequence, load limit, and travel path, and can be accurately adapted to various warehousing execution equipment such as AGVs, stacker cranes, conveyors, and palletizing robots.

[0036] The module possesses real-time status feedback and closed-loop management capabilities. During operation, it continuously collects parameters such as the operating speed, position, load, and working status of various execution devices, and synchronously transmits them back to the 3D digital twin model. This allows for real-time updates of the virtual scene's operation progress and equipment status, achieving dynamic synchronization between physical operations and virtual simulation. When issues such as equipment operating parameters exceeding preset thresholds, task execution lag, or path conflicts are detected, an early warning is immediately triggered, and a real-time rescheduling module is activated to update the scheduling scheme, ensuring continuous and stable operation of warehousing.

[0037] The present invention provides a method for intelligent warehousing and dynamic scheduling of waste household appliance raw materials based on digital twins. This method is implemented using the aforementioned intelligent warehousing and dynamic scheduling system and specifically includes the following steps: S1: Using 3D laser scanning technology combined with BIM modeling methods, a high-precision 3D digital twin model of waste household appliance raw material storage is built to complete the full-element mapping of storage physical space, raw material properties, and operating equipment, and to establish a real-time two-way data mapping channel between physical storage and virtual model.

[0038] S2: Through various sensor devices of the multi-source heterogeneous data sensing module, it collects raw material status data, equipment operation data and environmental monitoring data in the warehouse in real time around the clock, and updates them synchronously to the three-dimensional digital twin model at a preset frequency to realize the real-time update of the status of all elements of the warehouse.

[0039] S3: Based on the synchronized warehouse data from the digital twin model, accurately calculate the outbound priority of each raw material batch, specifically including the following sub-steps: S31: Retrieve core data such as category parameters, actual storage quantity, dismantling line capacity gap, order delivery deadline, and storage duration for all batches of raw materials to be shipped from the 3D digital twin model database; S32: Based on the raw material's metal content, environmental risk level, and resource reuse value, match a pre-set category weight table to determine the category weight coefficient for each batch of raw materials. By combining the real-time capacity gap of the dismantling line with the urgency of orders, a dismantling urgency coefficient is determined. ; S33: Calculate the normalized quantity value for each batch of raw materials based on the maximum storage capacity of the warehouse and the maximum storage threshold for raw materials. Normalized value of storage duration ; S34: Substitute into the multi-dimensional priority calculation formula to solve the outbound priority of each raw material batch. The raw material batches are sorted from highest to lowest priority value to determine the initial outbound order.

[0040] S4: Combining raw material outbound priority with the real-time warehouse status of the digital twin model, an improved adaptive genetic algorithm is used to solve the multi-objective scheduling optimization problem and generate the optimal initial scheduling scheme. This includes the following sub-steps: S41: Use integer encoding to uniformly encode all scheduled tasks and warehouse operation equipment to be executed, initialize the population size to 100, and generate the initial scheduling population; S42: Substitute the multi-objective fitness function, calculate the fitness value of all individuals in the population one by one, and select the best individual in the population; S43: The selection operation is carried out by combining the roulette wheel selection method with the elite retention strategy, and 10% of the high-fitness and high-quality individuals in the population are directly entered into the next generation of the population. S44: Based on the real-time fitness parameters of the population, adaptively adjust the crossover and mutation probabilities to complete the crossover and mutation operations of individuals in the population and update the population structure; S45: Determine whether the algorithm iteration count has reached the preset 200 times or the fitness value convergence error is less than 200 times. If any condition is met, the optimal scheduling scheme is output; otherwise, the iterative calculation step is returned for continuous optimization.

[0041] S5: Real-time monitoring of abnormal events throughout the entire warehousing operation process, completing anomaly classification and dynamic rescheduling, specifically including the following sub-steps: S51: Real-time collection of operational data from storage equipment, raw materials, and the environment; comparison with preset normal parameter thresholds; accurate identification of various abnormal events; and completion of level determination. S52: Based on the type and level of abnormal events, update the equipment status, raw material status, and task status data of the 3D digital twin model in real time, and synchronously mark abnormal operation nodes; S53: Match the anomaly level to determine the corresponding rolling window length, filter the core tasks participating in rescheduling based on the priority threshold, and remove low-priority non-urgent tasks. S54: Call the improved adaptive genetic algorithm to re-optimize and solve the tasks to be executed in the window, and generate an updated scheduling scheme that adapts to the current storage status; S55: Synchronize the updated scheduling scheme to the scheduling execution control module, and update the virtual scheduling status of the three-dimensional digital twin model to achieve virtual-real synchronous scheduling.

[0042] S6: The finalized scheduling plan is transformed into standardized equipment control commands, which are then sent to each warehouse execution equipment to complete automated storage and retrieval operations. The equipment execution status and operation progress are collected in real time throughout the process and continuously updated to the three-dimensional digital twin model to achieve closed-loop management of scheduling operations. Example 1: Initial Intelligent Scheduling of Warehouse under Normal Steady-State Operating Conditions

[0043] This embodiment is applied to a normal, steady-state operation scenario of undisturbed and abnormal storage of waste household appliance raw materials. The storage is equipped with 2 AGV handling devices and 1 automated stacker crane. The maximum storage capacity of raw materials in a single batch is... Taiwan, the maximum allowable storage time for waste household appliance raw materials Heaven. Preset weight parameters. , , , Scheduling target weight , , Energy consumption coefficient , There are currently 4 batches of waste household appliance raw materials awaiting shipment in the warehouse. The original parameters of each batch are as follows: Batch 1 is waste refrigerator raw material, category weight coefficient. The actual storage quantity is 30 units, and the urgency level of dismantling is [not specified]. The batch has been stored for 10 days; batch 2 consists of raw materials from waste washing machines, with a category weighting coefficient. The actual number of storage units is 25, and the urgency level of dismantling is [not specified]. The storage time has been 15 days; batch 3 is waste air conditioner raw material, category weight coefficient. The actual storage quantity is 40 units, and the urgency level of dismantling is [not specified]. The storage time has been 5 days; batch 4 is raw material from waste television sets, with a category weighting coefficient. The actual number of storage units is 20, and the urgency level of dismantling is [not specified]. It has been stored for 20 days.

[0044] First, the priority of raw material outbound shipment for each batch is calculated. The normalization parameters and priority calculation process are as follows: Batch 1: Quantity Normalized Value Storage duration normalized value Substituting into the formula, we get: .

[0045] Batch 2: Quantity Normalized Value Storage duration normalized value Substituting into the formula, we get: .

[0046] Batch 3: Quantity Normalized Value Storage duration normalized value Substituting into the formula, we get: .

[0047] Batch 4: Quantity Normalized Value Storage duration normalized value Substituting into the formula, we get: .

[0048] Based on the calculation results, the raw material outbound priority is ranked as batch 3, batch 1, batch 2, and batch 4. The priority data is imported into an improved adaptive genetic algorithm, the population is initialized, and iterative optimization is performed. The algorithm reaches the convergence threshold after 120 generations and outputs the optimal scheduling scheme. The total completion time for this scheduling is [not specified]. Total travel distance of the equipment Total energy consumption of equipment Substituting into the fitness function, we get: .

[0049] The final initial scheduling scheme was as follows: AGV1 prioritized executing the high-priority batch 3 outbound task, followed by batch 2 outbound operations. AGV2 then executed batch 1 and batch 4 outbound tasks sequentially. The stacker crane cooperated with the two AGVs throughout the entire process to complete the rack picking and palletizing operations. This scheduling scheme was imported into a digital twin model for virtual simulation verification. No equipment collisions, path conflicts, or operation deadlocks were found throughout the process. After the simulation verification passed, control commands were issued for execution, and the entire operation was completed in an orderly and efficient manner. Example 2: Dynamic Rescheduling in a Moderate Abnormal Scenarios with Single Device Failure

[0050] This embodiment, based on the initial scheduling scenario of Embodiment 1, simulates a moderate abnormal disturbance occurring during warehousing operations to verify the system's rescheduling capability. At the 60-minute mark of the initial scheduling plan, the system detects a sudden malfunction in AGV1 via the equipment status perception module. The walking mechanism jams and cannot continue performing the transportation task. Currently, AGV1 has completed all outbound operations for batch 3, with the remaining unfinished task being the outbound operation for batch 2. AGV2 is in normal operating condition and is currently performing the outbound operation for batch 1. The system automatically determines this malfunction to be a moderate abnormal event of single-equipment shutdown and matches the corresponding rolling window length for moderate abnormalities. Priority filtering threshold .

[0051] The system updates the status of the 3D digital twin model in real time, marks AGV1 as a fault-stopped state, locks the parameters of the remaining tasks to be executed, and filters the core tasks with a priority higher than 0.6 among the remaining tasks as batch 1. Batch 2 and batch 4 have a priority lower than the threshold and are not included in the scope of this rescheduling, maintaining their original standby scheduling state. Subsequently, the system automatically calls the improved adaptive genetic algorithm to re-optimize and solve the core tasks within the scrolling window.

[0052] After rescheduling optimization, the total completion time for this job is updated to: Total travel distance of the equipment Total energy consumption of equipment Substituting into the fitness function, we get: .

[0053] The generated optimal rescheduling scheme is as follows: suspend the regular operation of AGV2, prioritize the completion of the high-priority batch 1 outbound task, and avoid the risk of high-priority raw materials exceeding storage time limits; after AGV1 has been repaired and restored to normal operation, AGV1 will then take over to complete the regular outbound operations of batches 2 and 4. The updated scheduling scheme was simulated and verified in a digital twin model. No operation conflicts or equipment malfunctions were found, and the scheme was immediately deployed and executed after verification. This rescheduling only extended the overall operation time by 30 minutes, did not affect the dismantling and delivery nodes of high-priority waste household appliance raw materials, effectively avoided the environmental risks caused by long-term storage of raw materials, and achieved optimal scheduling adaptation under abnormal scenarios. Example 3: Dynamic Rescheduling of Urgent Orders in Severe Anomaly Scenarios

[0054] This embodiment, based on the initial steady-state scheduling scenario of Embodiment 1, simulates a severe anomaly caused by a sudden emergency environmental protection order in the warehouse, verifying the system's dynamic scheduling optimization capability under extreme scenarios. At the 30-minute mark of the initial scheduling plan, the system receives an emergency environmental disposal order from the superior dismantling platform, requiring priority completion of the outbound operations for three batches of highly polluting waste battery raw materials. These raw materials have a high environmental risk level and require timely disposal, classifying them as a top-level emergency task. The system automatically identifies this as a severe anomaly and matches the corresponding rolling window length to the severe anomaly. Priority filtering threshold .

[0055] The parameters for the three newly added batches of waste battery raw materials are as follows: Batch 5: Actual storage quantity 10 units, storage time 2 days, category weight coefficient. Disassembly urgency coefficient Batch 6 has an actual storage quantity of 15 units, a storage duration of 3 days, and a category weighting coefficient. Disassembly urgency coefficient Batch 7: Actual storage quantity 8 units, storage duration 1 day, category weight coefficient. Disassembly urgency coefficient The system updates the raw material database of the digital twin model in real time and recalculates the priority parameters of all batches to be shipped.

[0056] The process for calculating the priority of new batches is as follows: Batch 5: , , .

[0057] Batch 6: , , .

[0058] Batch 7: , , .

[0059] Based on the existing batch priority data, the updated priority order of all raw materials awaiting shipment is: Batch 3, Batch 6, Batch 5, Batch 7, Batch 1, Batch 2, and Batch 4. According to the threshold filtering rule, Batch 3, Batch 6, Batch 5, Batch 7, and Batch 1 have a priority higher than 0.6 and are included in the rescheduling optimization scope, while Batch 2 and Batch 4 have a priority lower than the threshold and their original scheduling order remains unchanged.

[0060] The system initiates a severe anomaly rescheduling optimization mechanism, using an improved adaptive genetic algorithm to iteratively solve for the optimal scheduling scheme, resulting in a reduced total job completion time. Total travel distance of the equipment Total energy consumption of equipment Substituting into the fitness function, we get: .

[0061] The final rescheduling plan is as follows: AGV1 continues to complete the existing high-priority batch 3 outbound operations, and immediately follows up with the emergency orders batches 6 and 5 after completion; AGV2 prioritizes completing the emergency order batch 7 operation, and then executes the existing high-priority batch 1 outbound operation; the regular outbound operations of batches 2 and 4 are postponed until all emergency environmental protection orders are completed. This plan has been verified by digital twin virtual simulation, showing reasonable operation sequence, no equipment conflicts, and sufficient guarantee of emergency task priority. It was immediately implemented after verification. This scheduling can complete the emergency outbound disposal of all highly polluting waste battery materials within 90 minutes, meeting the timeliness requirements of environmental emergency disposal, while minimizing the impact on regular warehousing operations, achieving optimal allocation of warehousing resources in emergency scenarios.

[0062] This invention achieves integrated and collaborative optimization of storage location management and outbound scheduling by constructing a full-element three-dimensional digital twin model of waste household appliance raw material storage and combining it with a multi-dimensional dynamic outbound priority calculation method. The digital twin model realizes real-time bidirectional mapping of physical storage space, raw material attributes, and equipment status, providing full-dimensional and high-precision status data support for scheduling decisions. Addressing the characteristics of the diverse categories and varying environmental risk levels of waste household appliance raw materials, a multi-dimensional priority evaluation system is established, including category weight, quantity, dismantling urgency, and storage duration. This system comprehensively considers the reuse value of raw materials, environmental disposal requirements, and dismantling production needs, avoiding problems such as excessive storage time for highly polluting raw materials and delays in core dismantling orders. It also overcomes the shortcomings of existing technologies that only focus on inbound storage location allocation and lack outbound scheduling optimization.

[0063] This invention employs an improved adaptive genetic algorithm to solve the multi-objective optimization problem of warehouse scheduling. By introducing adaptive crossover and mutation probabilities, and dynamically adjusting algorithm parameters based on the real-time fitness of the population, it effectively solves the problems of premature convergence and low search efficiency in the later stages of traditional genetic algorithms, thus improving the solution accuracy and convergence speed of multi-objective optimization problems. The algorithm constructs a multi-objective fitness function that includes total scheduling completion time, total equipment travel distance, and total equipment energy consumption. This enables the reduction of energy consumption and equipment wear in warehouse operations while ensuring scheduling efficiency, achieving multi-objective optimal allocation of warehouse resources and overcoming the limitations of existing technologies using fixed-parameter genetic algorithms.

[0064] This invention establishes a graded response and dynamic rolling time-domain rescheduling mechanism for abnormal events. Warehouse abnormal events are classified into three levels: minor, moderate, and severe. The rolling window length is dynamically adjusted for different levels of abnormality, and a task priority filtering mechanism is used to select the scope of rescheduling tasks. This mechanism can respond quickly when abnormal events occur, adjust the scheduling plan in a timely manner to ensure the continuity of warehouse operations, and effectively reduce the computational load of rescheduling by limiting the scope of rescheduling tasks, thereby improving the scheduling response speed. It solves the problems of existing technologies lacking a comprehensive abnormality handling mechanism and easily causing operation interruptions when abnormalities occur.

[0065] This invention achieves closed-loop management of the warehouse scheduling process through multi-source heterogeneous data fusion technology and virtual simulation verification. The multi-source heterogeneous data sensing module uses DS evidence theory to fuse data from multiple sensors, improving the accuracy of raw material identification and status monitoring. Virtual simulation verification is performed in a digital twin model before the scheduling plan is executed, enabling early detection of issues such as equipment collisions, path conflicts, and operational deadlocks, thus avoiding ineffective scheduling and operational accidents. During scheduling execution, real-time feedback on equipment operating status and task progress is provided, achieving dynamic synchronization between physical operations and the virtual model. This improves the reliability and stability of the warehouse scheduling system, meeting the practical needs of large-scale and intelligent operation of waste household appliance raw material warehousing.

[0066] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins, characterized in that, include: The digital twin warehouse modeling module is used to construct a three-dimensional digital twin model of waste household appliance raw material warehouse. The three-dimensional digital twin model includes a warehouse physical space mapping unit, a raw material attribute mapping unit, a warehouse equipment mapping unit, and a virtual simulation verification unit. The multi-source heterogeneous data sensing module is used to collect raw material status data, equipment operation data and environmental data in the warehouse in real time, and synchronize the collected data to the three-dimensional digital twin model at a preset frequency; The dynamic priority calculation module is used to calculate the outbound priority of each batch of raw materials based on the type, quantity, urgency of dismantling, and storage duration of the raw materials using a multi-dimensional priority calculation formula. The hybrid intelligent scheduling module is used to solve the multi-objective optimization problem of warehouse scheduling based on real-time status data of the three-dimensional digital twin model and the outbound priority of each batch of raw materials, and to generate an initial scheduling scheme by using an improved adaptive genetic algorithm. The real-time rescheduling trigger module is used to monitor various abnormal events in the warehouse. When an abnormal event occurs, it triggers the corresponding level of real-time rescheduling process according to the severity of the abnormal event. The scheduling execution control module is used to convert scheduling schemes into standardized control commands, which are then sent to execution equipment such as AGVs, stacker cranes, conveyors, and palletizing robots in the warehouse, and the execution status is fed back to the three-dimensional digital twin model in real time.

2. The dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 1, characterized in that, The multi-dimensional priority calculation formula in the dynamic priority calculation module is as follows: ； in, For the first The outbound priority of each raw material batch, with a value range of [value range missing]. ; For the first The category weighting coefficient for each batch of raw materials is determined based on the metal content, environmental risk level, and reuse value of the raw materials, with a range of values ​​ranging from [value missing]. ; For the first The normalized quantity value of each raw material batch is calculated using the following formula: , This represents the maximum storage capacity per batch. For the first The urgency coefficient for dismantling each batch of raw materials is determined based on the real-time capacity gap of the dismantling line and the order delivery deadline, with a value range of [value missing]. ; For the first The normalized value of the storage duration for each batch of raw materials is calculated using the following formula: , This refers to the maximum permissible storage time for raw materials; Let be the weight coefficient, and satisfy... .

3. The dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 1, characterized in that, The improved adaptive genetic algorithm used in the hybrid intelligent scheduling module has a multi-objective fitness function as follows: ； in, This is the fitness value; The total completion time for all scheduled tasks; The total travel distance of all executing devices; The total energy consumption of all executing devices is calculated using the following formula: , The total number of devices to be executed. For the first The load capacity of the equipment For the first The travel distance of the equipment. Energy consumption coefficient per unit load per unit distance The standby power consumption coefficient per unit time. For the first Standby time of the device; Let be the target weight coefficient, and satisfy... .

4. The dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 3, characterized in that, The improved adaptive genetic algorithm introduces both adaptive crossover probability and adaptive mutation probability, where the adaptive mutation probability is calculated using the following formula: ； in, For adaptive mutation probability; The maximum mutation probability is given by a value of [value]. ; To minimize the mutation probability, the value is [value]. ; This represents the average fitness value of the current population. The fitness value of the individual to be mutated; This represents the minimum fitness value of the current population.

5. The dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 1, characterized in that, The real-time rescheduling triggering module adopts a dynamic scrolling time-domain optimization strategy, setting the basic scrolling window length to [value missing]. When an abnormal event occurs, it is first classified into three levels—minor, moderate, and severe—according to a preset abnormality classification standard. Then, the length of the scrolling window is dynamically adjusted: for minor abnormalities... When moderate abnormality When there is a serious abnormality Simultaneously, a task priority filtering mechanism is introduced, only processing tasks with an outbound priority higher than a preset threshold. The tasks are re-optimized, while tasks outside the rolling window and with a priority below the threshold retain their original scheduling scheme.

6. The dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 1, characterized in that, The multi-source heterogeneous data sensing module includes an RFID tag reader, a 3D vision sensor, a weight sensor, a temperature and humidity sensor, and a smoke sensor. This module employs DS evidence theory to fuse raw material data collected by the RFID tag reader, 3D vision sensor, and weight sensor to improve raw material identification accuracy. The data synchronization frequency is set to: equipment operating status data and raw material location data every [time / period]. Synchronize once, environmental data every Once synchronized, raw material attribute data is synchronized in real time during both inbound and outbound processes.

7. A dynamic scheduling method for intelligent warehousing of waste household appliance raw materials based on digital twins, characterized in that, This method is implemented based on the dynamic scheduling system for intelligent warehousing of waste household appliance raw materials based on digital twins as described in any one of claims 1-6, and includes the following steps: S1: Construct a three-dimensional digital twin model of the waste household appliance raw material storage and establish a real-time data mapping channel between the physical storage and the digital twin model; S2: Real-time collection of raw material status data, equipment operation data, and environmental data within the warehouse via a multi-source heterogeneous data sensing module, and updating of these data to the three-dimensional digital twin model; S3: Calculate the outbound priority of each batch of raw materials based on the type, quantity, urgency of dismantling, and storage duration using a multi-dimensional priority calculation formula; S4: Based on the real-time status data of the three-dimensional digital twin model and the outbound priority of each batch of raw materials, an improved adaptive genetic algorithm is used to solve the multi-objective optimization problem of warehouse scheduling and generate an initial scheduling scheme. S5: Monitors various abnormal events in the warehouse. When an abnormal event occurs, it triggers the corresponding level of real-time rescheduling process according to the severity of the abnormal event and generates an updated scheduling plan. S6: Convert the scheduling scheme into standardized control commands, issue them to the execution equipment in the warehouse, and provide real-time feedback on the execution status to the three-dimensional digital twin model.

8. The dynamic scheduling method for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 7, characterized in that, Step S3 specifically includes the following steps: S31: Obtain the category information, actual quantity, dismantling line capacity gap data, and stored duration data of all raw material batches to be shipped from the three-dimensional digital twin model; S32: Determine the category weight coefficient for each batch of raw materials based on the preset category weight table. The urgency coefficient for dismantling each batch of raw materials is determined based on the dismantling line capacity gap and order delivery deadlines. ; S33: Calculate the normalized quantity values ​​for each batch of raw materials. and storage duration normalized value ; S34: Substitute into the multi-dimensional priority calculation formula to calculate the outbound priority of each raw material batch. The raw material batches are sorted from highest to lowest priority.

9. The dynamic scheduling method for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 7, characterized in that, Step S4 specifically includes the following steps: S41: Encode the scheduling task and execution device using integer encoding to generate an initial population, with the population size set to [value missing]. ; S42: Calculate the fitness value of each individual in the population based on the multi-objective fitness function; S43: Use a roulette wheel selection method combined with an elite retention strategy to select the option with the highest fitness value. Individuals directly enter the next generation; S44: Perform crossover and mutation operations on individuals in the population based on adaptive crossover and adaptive mutation probabilities; S45: Determine if the preset number of iterations has been reached. Or fitness value convergence threshold If so, output the optimal scheduling scheme; otherwise, return to step S42 to continue iterating.

10. The dynamic scheduling method for intelligent warehousing of waste household appliance raw materials based on digital twins according to claim 7, characterized in that, Step S5 specifically includes the following steps: S51: Real-time monitoring of abnormal events within the warehouse; when an abnormal event is detected, it is classified and its level is determined. S52: Update the corresponding status data of the 3D digital twin model according to the type and level of the abnormal event, including equipment failure status, raw material damage status and task change status; S53: Determine the length of the rolling window and the scope of tasks participating in rescheduling based on the level of the abnormal event, and filter out tasks with priority below the threshold; S54: Call the improved adaptive genetic algorithm to re-optimize and solve the tasks within the rolling window, and generate an updated scheduling scheme; S55: The updated scheduling scheme is sent to the scheduling execution control module, and the virtual scheduling status in the three-dimensional digital twin model is updated at the same time.

Images