An improved data twin generation method and system based on an ant colony algorithm

Abstract

The application discloses an improved data twin generation method and system based on an ant colony algorithm, and the method comprises the following steps: associating a task-executing forklift with a current task; in response to a task-executing forklift route planning request, determining a starting point and a target point of the task-executing forklift; based on the relative positions of the starting point and the target point, generating a plurality of feasible routes, and putting the plurality of feasible routes into a feasible route dataset; calculating the total generation value Yn of the feasible routes, determining the route with the minimum total generation value Yn as the latest path of the forklift, and putting the latest path of the forklift into a warehouse forklift latest path dataset; monitoring the paths in the warehouse forklift latest path dataset two by two, judging whether there is a collision conflict, if yes, determining the collision type, calculating the avoidance cost, and determining the avoidance forklift and the avoidance strategy corresponding to the avoidance forklift based on the collision type and the avoidance cost; and the technical scheme of the application can solve the problem that the route planning in the prior art is not scientific enough.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a data twin generation method and system based on an improved ant colony algorithm. Background Technology

[0002] In modern intelligent warehouse management systems, data twin technology plays an increasingly crucial role, aiming to build accurate virtual models for warehouse operations to achieve efficient control. However, many current related technical solutions still have significant shortcomings.

[0003] On the one hand, in the forklift task allocation process, traditional methods often lack detailed consideration. Most systems simply assign tasks based on whether the forklift is idle, without paying in-depth attention to the idle time of the forklift and the update time of the status information. This can easily lead to some forklifts that have been idle for a long time and are in urgent need of new tasks being ignored, resulting in idle resources and waste, and reducing overall operational efficiency. Especially during peak periods of goods entering and leaving the warehouse, the contradiction between task backlog and forklift idleness is more prominent, which cannot meet the needs of rapid turnover in warehousing and logistics.

[0004] On the other hand, forklift route planning is not scientific and perfect enough. Most existing technologies focus solely on travel distance and try to find the shortest path, but ignore the important impact of turning operations on forklift operating efficiency, energy consumption and cargo stability. Frequent turning not only increases mechanical wear and tear on forklifts and consumes more energy, but may also endanger cargo safety due to inertia. Therefore, simply pursuing the shortest travel distance is not the optimal choice, affecting the smoothness and accuracy of warehousing operations.

[0005] Furthermore, there is a serious lack of handling forklift collisions along their routes. As warehouse automation increases, multiple forklifts are operating simultaneously more frequently, leading to a surge in the risk of collisions along their paths. However, current systems generally lack real-time and effective collision monitoring mechanisms. Even when a collision is detected, it is difficult to quickly and accurately determine the type of collision, let alone quantify the cost of avoidance. Consequently, it is difficult to develop targeted and cost-effective avoidance strategies for forklifts, often resulting in operational stoppages, cargo damage, and other problems, causing numerous problems and economic losses for warehouse operations.

[0006] In summary, there is an urgent need for an innovative data twin generation method based on an improved ant colony algorithm to comprehensively optimize key aspects such as forklift task allocation, route planning, and collision and conflict handling, so as to improve the overall efficiency of intelligent warehouse management. Summary of the Invention

[0007] In view of this, the present invention proposes a data twin generation method and system based on an improved ant colony algorithm, which can improve forklift operation efficiency, ensure operational safety, optimize the overall efficiency of warehouse management, and help the intelligent warehousing system operate efficiently by rationally allocating forklift tasks, comprehensively considering route planning factors, and effectively monitoring and responding to path collisions.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A data twin generation method based on an improved ant colony algorithm includes:

[0010] In response to a task request, the status information of all forklifts is obtained, and the forklift with the idle status information and the earliest update time is selected to execute the task. The forklift executing the task is associated with the current task, wherein the task includes goods inbound task, goods transfer task, and goods outbound task.

[0011] In response to a forklift route planning request for a task, determine the starting point and destination point of the forklift.

[0012] Based on the relative positions of the starting point and the target point, multiple feasible routes are generated and placed into a feasible route dataset.

[0013] Calculate the total driving cost Yn of the feasible route, where the total driving cost Yn = driving cost Sn + turning cost Zn. Determine the route with the minimum total driving cost Yn as the latest forklift path and add the latest forklift path to the warehouse forklift latest path dataset.

[0014] Monitor the pairs of paths in the latest path dataset of warehouse forklifts to determine if there is a collision conflict. If so, determine the collision type and calculate the avoidance cost. Based on the collision type and avoidance cost, determine the forklift to avoid and the avoidance strategy corresponding to the forklift to avoid.

[0015] Based on the above technical solution, the present invention can be further improved as follows:

[0016] Furthermore, the data twin generation method based on the ant colony algorithm also includes:

[0017] Project the warehouse digital raster map onto the raster matrix so that the forklift's movement step is equal to the side length of the raster.

[0018] Mark the grid coordinates corresponding to the lanes as feasible areas, mark the grid coordinates corresponding to the parking spaces as infeasible areas, and mark the grid coordinates between the lanes and parking spaces as infeasible areas to construct a grid matrix.

[0019] Initialize the pheromone matrix Tau, and set the maximum number of iterations K, the number of ants M, the pheromone importance parameter, the heuristic factor importance parameter, the pheromone evaporation coefficient, and the pheromone increase intensity coefficient;

[0020] Establish a matrix D to store the cost of each grid point to its adjacent unobstructed grid points;

[0021] Construct a heuristic information matrix, using the reciprocal of the distance from the current grid point to the target grid point as the heuristic information, and setting the heuristic information at obstacles to 0.

[0022] Furthermore, based on the relative positions of the starting point and the target point, multiple feasible routes are generated, and these multiple feasible routes are placed into a feasible route dataset, including:

[0023] For each ant, initialize the ant's crawling path and path length. The ant starts from the starting point and begins to search for the path, finding all unobstructed adjacent grid points of the current grid point, and filtering the available grid points based on the taboo list.

[0024] If the current grid point is not the target grid point and the number of selectable grid points is greater than or equal to 1, calculate the probability that the ant will move from the current grid point to each of the adjacent grid points, select the next grid point based on the roulette wheel method, update the ant's crawling path, path length, matrix D and taboo list, until the current grid point is the target grid point or the number of selectable grid points is less than 1, end the current iteration, and record the ant's walking route;

[0025] If the ant's last step is the target grid point, then calculate the current path length and compare it with the currently known shortest path length. If the current path length is less than the currently known shortest path length, then update the currently known shortest path length and the shortest path.

[0026] If the ant's last step is not the target grid point, then the path length is recorded as 0;

[0027] Repeat the path search steps until M ants have completed one round of path search. Determine if the termination condition is met. If so, end the ant colony algorithm and draw the optimal planned path.

[0028] Furthermore, the collision types include straight-on collisions, collisions expected to occur on different routes, perpendicular collisions, and chase collisions:

[0029] When the collision type is a straight-on collision:

[0030] If a heavily loaded forklift encounters an unloaded forklift, the route for the unloaded forklift will be replanned at a step E away from the point of collision.

[0031] If the forklifts are under the same load, the forklift with the lower avoidance cost should re-plan its route at a step distance E or F from the collision point.

[0032] If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will re-plan its route at a step length E or F away from the collision point.

[0033] When the collision type is a predicted head-on collision along different routes:

[0034] If a heavily loaded forklift encounters an unloaded forklift, the unloaded forklift will wait to avoid the collision at a distance G from the point of impact.

[0035] If the forklifts are under the same load, calculate the collision location, and the forklift with the lower avoidance cost will wait to avoid the collision at a step length G or H away from the collision point.

[0036] If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will wait to avoid the collision at a step length G or H set at the point of collision.

[0037] When the collision type is vertical collision:

[0038] If a heavily loaded forklift encounters an unloaded forklift, the unloaded forklift will wait to avoid the collision at a distance M from the point of impact.

[0039] If the forklifts are under the same load, calculate the collision location and have the forklift with the lower avoidance cost wait to avoid the collision at a step distance M or N from the collision point.

[0040] If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will wait to avoid the collision at a step distance M or N from the collision point.

[0041] When the collision type is chasing collision:

[0042] When the following vehicle yields, it maintains a preset step length of I or J. After the preceding vehicle turns and moves away from the preset distance, it continues to drive normally.

[0043] Furthermore, the data twin generation method based on the ant colony algorithm also includes:

[0044] Based on actual warehouse operations, we define 0 as the forklift idle state, 1 as the forklift completing its task and heading to the waiting area, 2 as the forklift empty and heading to the inbound loading area, 3 as the forklift heavily loaded and heading to the inbound storage location, 4 as the forklift empty and heading to the outbound storage location, 5 as the forklift in the outbound process, 6 as the forklift heavily loaded and heading to the unloading area, 7 as the forklift empty and heading to the transfer outbound storage location, 8 as the forklift in the transfer process, and 9 as the forklift heavily loaded and heading to the transfer inbound storage location.

[0045] Define 00 as goods awaiting warehousing, 11 as goods in warehousing and transportation status, 22 as goods already in warehousing status, 33 as goods awaiting transfer to warehousing status, 44 as goods transferred and removed from shelves status, 55 as goods in warehousing and transportation status, 66 as other goods transferred and removed from shelves status, 77 as goods awaiting shipment status, 88 as goods shipped and removed from shelves status, 99 as goods shipped and transportation status, 100 as goods already shipped status, and 110 as other goods shipped and removed from shelves status.

[0046] The status flow for forklift inbound operations is 0-2-3-1-0, the status flow for forklift outbound operations is 0-4-5-6-1-0, and the status flow for forklift transfer operations is 0-7-8-9-1-0.

[0047] Furthermore, the data twin generation method based on the ant colony algorithm also includes:

[0048] Receive the inbound business order order1 (O1, O2, ... On), associate the goods An with the business order On, and set the status of goods An to 00;

[0049] Get the forklift with a status of 0, assign a forklift Cn, and associate the business order On with the forklift Cn;

[0050] Based on all the data of warehouse locations that are not full, a recommended location Kx is obtained. Based on the recommended location Kx, the starting point and target point of the forklift to perform the task are determined. Based on the starting point and target point, a route is planned for the forklift.

[0051] Forklift Cn goes to the pickup area to scan the tag code of goods An and associates it with goods An. The status of goods An is 11.

[0052] After the forklift Cn transports the goods to the warehouse location, the label code of goods An is scanned a second time to confirm that goods An has been put on the shelf.

[0053] The business scheduling platform receives the scanned information and confirms the actual storage location Kn where the forklift is placed.

[0054] If the business scheduling platform obtains the number N of goods in storage location Kn with a status of 22, then N+1 is the layer Hn of goods An in storage location Kn with a status of 22.

[0055] The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to create a digital twin model of the goods.

[0056] Furthermore, the data twin generation method based on the ant colony algorithm also includes:

[0057] Receive the transfer order order2(O1, O2, ... On), associate the goods An with the order On, and set the status of goods An to 33;

[0058] Get the forklift with a status of 0, assign a forklift Cn, and associate the current order On with the forklift Cn;

[0059] Obtain the current storage location K of goods An and the target storage location K1 for relocation. Based on the target storage location K1 and the current storage location K, determine the starting point and target point of the forklift to perform the task, and plan the route for the forklift based on the starting point and target point.

[0060] The business scheduling platform obtains the layer number Fn of goods An in the current storage location K and the total layer number F of goods in the current storage location. F-Fn represents the goods (Ay1, Ay2, ...) above goods An.

[0061] Determine if the number of layers Fn is equal to the total number of layers F. If so, the forklift scans the barcode to remove goods An from the shelf. The goods status of goods An is 55. The forklift moves goods An to the target storage location K1, puts it on the shelf, and scans the barcode of goods An. The business scheduling platform receives the barcode information and confirms the actual storage location Kn where the forklift puts it on. The business scheduling platform obtains the number N of goods in storage location Kn with a status of 22. Then N+1 is the number of layers Hn of goods An in storage location Kn with a status of 22. The data platform generates the storage location and corresponding layer number of each goods in the warehouse in real time and generates a digital twin model of the goods.

[0062] If not, the forklift scans the barcode and gradually removes the goods (Ay1, Ay2...) above Goods An. The status of the goods (Ay1, Ay2...) above Goods An is 66. The forklift scans the barcode and removes Goods An. The status of Goods An is 44. The forklift scans the barcode and gradually puts the goods (Ay1, Ay2...) above Goods An back on. The layer number of the goods (Ay1, Ay2...) above Goods An is the original layer number minus 1, and the status is 22. The forklift scans the barcode... Goods An has a status of 55. A forklift moves to the target storage location K1, puts the goods An on the shelf, and scans the barcode. The business scheduling platform receives the barcode information and confirms the actual storage location Kn where the forklift puts the goods on the shelf. The business scheduling platform obtains the number N of goods in storage location Kn with a status of 22. Then N+1 is the layer Hn of goods An in storage location Kn. The status of goods An is 22. The data platform generates the storage location and corresponding layer number of each goods in the warehouse in real time, generating a digital twin model of the goods.

[0063] Furthermore, the data twin generation method based on the ant colony algorithm also includes:

[0064] Receive a request for goods An to be shipped out, associate the goods An with the transfer order order3 (O1, O2, ... On), associate the current order On with the current goods An, and the goods status is 77;

[0065] Get the forklift with a status of 0, assign a forklift Cn, and associate the current order On with the forklift Cn;

[0066] Obtain the current storage location K of goods An, determine the starting point and target point of the forklift to perform the task based on the current storage location K, and plan the route for the forklift based on the starting point and target point;

[0067] The business scheduling platform obtains the layer number Fn of goods An in storage location K and the total layer number F of goods in the current storage location, where F-Fn is the goods (Ay1, Ay2, ...) above goods An.

[0068] Determine if the number of layers Fn is equal to the total number of layers F. If so, the forklift scans the barcode of goods An and removes them from the shelf, setting the status of goods An to 99. The forklift then proceeds to the outbound area, completes the outbound process, and scans the barcode of goods An. The business scheduling platform receives the barcode information and confirms that the forklift is in the outbound picking area. The business scheduling platform marks the status of goods An as 100. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to generate a digital twin model of the goods.

[0069] If not, the forklift scans the goods (Ay1, Ay2...) above goods An and gradually removes them from the shelves, making the goods status of goods (Ay1, Ay2...) above goods An 110. The forklift scans the goods An to remove them from the shelves, making the goods status of goods An 88. The forklift scans the goods (Ay1, Ay2...) above goods An and gradually puts them back on the shelves, making the goods status of goods (Ay1, Ay2...) above goods An 22. The layer number is the original record layer number minus 1, and the status is 22. The forklift scans goods An, and the goods status of goods An is 99. The forklift moves to the outbound area to complete the outbound process and scans goods An. The business scheduling platform receives the scanning information and confirms that the forklift is in the outbound picking area. The business scheduling platform marks the status of goods An as 100. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to generate a digital twin model of the goods.

[0070] A data twin generation system based on an improved ant colony algorithm includes:

[0071] The forklift task allocation module is used to respond to task requests, obtain the status information of all forklifts, select the forklift whose status information is idle and whose update time is the earliest to execute the task, and associate the forklift executing the task with the current task order. The tasks include goods receiving tasks, goods transfer tasks, and goods outgoing tasks.

[0072] The path planning request and response module is used to respond to the forklift route planning request for performing the task and determine the starting point and target point of the forklift performing the task.

[0073] The feasible route generation module is used to generate multiple feasible routes based on the relative positions of the starting point and the target point, and put the multiple feasible routes into the feasible route dataset;

[0074] The optimal route determination module is used to calculate the total travel cost Yn of feasible routes, determine the route with the minimum total travel cost Yn as the latest route of the forklift, and put the latest route into the warehouse forklift latest route dataset.

[0075] The collision detection and avoidance module is used to monitor the pairs of paths in the latest path dataset of warehouse forklifts, determine whether a collision has occurred, and if so, determine the collision type and calculate the avoidance cost. Based on the dynamic priority avoidance strategy, it obtains the status of the forklifts to avoid the collision and determines whether the forklifts need to replan their routes. If so, it dynamically plans the path with the minimum operating cost for the forklifts until the minimum cost path planning for all forklifts is completed.

[0076] An electronic device includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method described herein.

[0077] The present invention has the following advantages:

[0078] This invention presents a data twin generation method based on an improved ant colony algorithm. In forklift task allocation, by accurately acquiring forklift status information, it prioritizes the use of idle forklifts with the earliest update time, significantly improving forklift resource utilization and ensuring timely response to tasks such as goods warehousing, transfer, and outbound, reducing task backlog and waiting time. At the route planning level, it comprehensively considers the starting and ending points to generate multiple routes and innovatively combines travel and turning costs to calculate the total travel value, thereby selecting the optimal forklift path. This not only shortens transportation distances but also reduces forklift energy consumption, decreases mechanical wear, improves operational stability, and ensures safe goods transportation. For multi-forklift operation scenarios, it can monitor collisions between forklifts along their latest paths, promptly determine collision types, calculate avoidance costs, and rationally determine the forklifts to avoid and their avoidance strategies. This effectively avoids operational stoppages and the risk of goods damage, comprehensively improving the smoothness, efficiency, and safety of warehousing operations, and injecting powerful momentum into intelligent warehouse management.

[0079] The data twin generation method based on the improved ant colony algorithm in this invention defines various business states of forklifts (such as inbound, outbound, and transfer states) and various states of goods. After receiving inbound, outbound, and transfer business orders, it can accurately allocate idle forklifts, plan forklift routes based on warehouse location data, and orderly handle operations such as unloading, shelving, and transportation of goods. It can also obtain real-time information on the location and layer of goods and efficiently generate a digital twin model of goods that reflects the real-time situation of the warehouse, thereby realizing intelligent, visual, and precise warehouse management. Attached Figure Description

[0080] For illustrative and not limiting purposes, the present invention will now be described in conjunction with embodiments and accompanying drawings, wherein:

[0081] Figure 1 This is a flowchart illustrating the data twin generation method based on the improved ant colony algorithm in an embodiment of the present invention.

[0082] Figure 2 This is a schematic diagram of the main components of the data twin generation system based on the improved ant colony algorithm in an embodiment of the present invention;

[0083] Figure 3 This is a schematic diagram of a grid diagram in an embodiment of the present invention;

[0084] Figure 4 This is a schematic diagram showing the trend of the convergence curve of the first experimental result in the embodiment of the present invention.

[0085] Figure 5 This is a schematic diagram of the path trajectory of the first experimental result in an embodiment of the present invention;

[0086] Figure 6 This is a schematic diagram showing the trend of the convergence curve of the second experimental result in the embodiment of the present invention;

[0087] Figure 7 This is a schematic diagram of the path trajectory of the second experimental result in an embodiment of the present invention;

[0088] Figure 8 This is a schematic diagram of the physical structure of the electronic device provided by the present invention. Detailed Implementation

[0089] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0090] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate for the embodiments of the invention described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0091] It should be noted that, where there is no conflict, the embodiments and features of the present invention can be combined with each other. The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0092] Figure 1 This is a flowchart illustrating the data twin generation method based on the ant colony algorithm improved in this embodiment of the invention, as shown below. Figure 1 As shown, the data twin generation method based on the improved ant colony algorithm provided in this embodiment of the invention includes the following steps S101 to S105.

[0093] S101, in response to the task request, obtain the status information of all forklifts, select the forklift whose status information is idle and whose update time is the earliest to execute the task, and associate the forklift executing the task with the current task.

[0094] Specifically, the tasks include goods receiving, goods transferring, and goods leaving the warehouse.

[0095] The system works in real-time with the warehouse management system, sending data requests to the forklift management module at regular intervals (10 seconds) to collect detailed status information for each forklift. This information includes the forklift's real-time location, battery or fuel level, load status, mechanical component operating status, and the time of the last task completion. Timestamps are used to mark the update time of each forklift's status information. Through an intelligent filtering algorithm, the system quickly and accurately locates the forklift that is both idle and has the earliest update time among the idle forklifts. Once identified, the system immediately establishes a connection between this forklift and the current pending task (goods receiving, goods transfer, or goods dispatch; task details also come from the warehouse management system's task allocation pool, including key information such as goods type, quantity, starting storage location, and target storage location). The task details are then pushed to the selected forklift's onboard terminal, ensuring that the forklift driver is aware of the task and can start work immediately, avoiding idle forklift resources and improving the overall response speed of warehouse operations.

[0096] S102, in response to the forklift route planning request, determines the starting point and target point of the forklift performing the task.

[0097] Specifically, when a forklift receives a task instruction, its onboard terminal automatically sends a route planning request to the data twin system. Based on the task objective, it accurately identifies the destination to which the goods need to be delivered, such as the specific location of a row of shelves in the warehouse, the exit or the inbound platform, and sets it as the target point.

[0098] The improved data twin generation method based on the ant colony algorithm generates multiple feasible routes based on the starting point and the target point, and determines the optimal path by comprehensively considering the travel cost and the turning cost. This method not only shortens the forklift travel distance and reduces energy consumption, but also takes into account the impact of forklift turning operations on the stability of goods and mechanical wear, ensuring safe and efficient goods transportation and optimizing warehousing operations.

[0099] S103: Based on the relative positions of the starting point and the target point, generate multiple feasible routes and put these feasible routes into the feasible route dataset.

[0100] Specifically, such as Figure 3 As shown, using the grid method, the map is abstracted as a 0-1 matrix. Each grid cell represents whether two adjacent cells can be safely traversed, indicated by a value of 0 or 1. A value of 0 indicates that two cells are completely infeasible, while a value of 1 indicates that they are completely feasible. The cost of the grid matrix is ​​then calculated. Using the ant colony algorithm, the path of each ant is recorded. The next grid cell is selected using a roulette wheel approach, and the ant's path, path length, and taboo list are updated. This process is repeated until the Mth ant completes a round of path searching. The pheromone is then updated according to a formula, and the termination condition is checked. Finally, the ant search path map is drawn, which represents the forklift's travel route.

[0101] This invention projects a warehouse digital grid map onto the aforementioned grid matrix, where the forklift's movement step is equal to the grid's side length. Based on the grid coordinates corresponding to lanes as feasible areas, storage locations as infeasible areas, and the area between lanes and storage locations as infeasible areas, the grid matrix is ​​established as follows: fully feasible grids are represented as 0, and fully infeasible grids as 1. The entire environment divided by the grid is numbered (this invention uses a numbering method). The grid method uses each grid as a unit.

[0102] The data twin generation method based on the improved ant colony algorithm projects the warehouse digital raster map into a raster matrix, clearly delineates feasible and infeasible areas, accurately simulates the actual warehouse layout, provides an accurate digital environment foundation for forklift path planning, makes path planning more realistic, and reduces path infeasibility problems caused by map errors.

[0103] Based on the above grid matrix, the present invention establishes the following ant colony algorithm experiment:

[0104] Initialize the pheromone matrix Tau, and set the maximum number of iterations K, the number of ants M, the pheromone importance parameter, the heuristic factor importance parameter, the pheromone evaporation coefficient, and the pheromone increase intensity coefficient;

[0105] Establish a matrix D to store the cost of each grid point to its adjacent unobstructed grid points;

[0106] Construct a heuristic information matrix, using the reciprocal of the distance from the current grid point to the target grid point as the heuristic information, and setting the heuristic information at obstacles to 0.

[0107] The data twin generation method based on the improved ant colony algorithm reasonably sets various parameters of the ant colony algorithm, such as the maximum number of iterations, the number of ants, and the pheromone correlation coefficient. Combined with the constructed pheromone matrix, cost matrix, and heuristic information matrix, the algorithm can quickly converge to a better solution when searching for forklift paths, thereby improving path planning efficiency and reducing computational resource consumption.

[0108] For each ant, initialize the ant's crawling path and path length. The ant starts from the starting point and begins to search for the path, finding all unobstructed adjacent grid points of the current grid point, and filtering the available grid points based on the taboo list.

[0109] If the current grid point is not the target grid point and the number of selectable grid points is greater than or equal to 1, calculate the probability that the ant will move from the current grid point to each of the adjacent grid points, select the next grid point based on the roulette wheel method, update the ant's crawling path, path length, matrix D and taboo list, until the current grid point is the target grid point or the number of selectable grid points is less than 1, end the current iteration, and record the ant's walking route;

[0110] If the ant's last step is the target grid point, then calculate the current path length and compare it with the currently known shortest path length. If the current path length is less than the currently known shortest path length, then update the currently known shortest path length and the shortest path.

[0111] If the ant's last step is not the target grid point, then the path length is recorded as 0;

[0112] Repeat the path search steps until M ants have completed one round of path search. Determine if the termination condition is met. If so, end the ant colony algorithm and draw the optimal planned path.

[0113] The data twin generation method based on the improved ant colony algorithm utilizes multiple ants to search in parallel. It selects paths from numerous optional grid points based on a tabu list and roulette wheel method, fully simulating the collective intelligent behavior of ants foraging. In complex warehouse grid map environments, it efficiently explores multiple feasible routes, increasing the probability of finding the globally optimal path, which is more advantageous than traditional single search methods. During the search process, the path length and matrix information are updated in real time and compared with the known shortest path, continuously iterating and optimizing the path to ensure that the final optimal planned path can adapt to complex situations such as dynamic changes in warehouse layout and diverse task requirements, effectively improving the actual operating efficiency of forklifts.

[0114] like Figures 4-7 As shown, the following are the test results of this invention based on the ant colony algorithm in different environments;

[0115] First experimental result: The total path length is S = 29.2132034355964;

[0116] Second experimental result: The total path length is S = 16.7279220613579;

[0117] This invention provides a global route planning method based on an improvement of the ant colony algorithm, which plans the shortest path for forklifts in a dynamic random network. Global route planning includes global static route planning and dynamic local route planning. Global static route planning comprehensively considers the current grid map, forklift status information, task information, etc., to pre-plan the paths for all forklifts, thereby obtaining the globally optimal solution, which has strong anti-interference capabilities. However, as the warehouse environment changes and the number of forklifts increases, the computational load of global static route planning becomes increasingly large, making it difficult to efficiently calculate the optimal solution. Therefore, this invention combines the global static route planning method with the dynamic local route planning method. When conflicts occur in warehouse forklift transportation, it can prioritize planning a separate shortest avoidance path for each forklift, achieving coordinated control of multiple forklifts and real-time planning of conflict-free paths, thus solving the problems of efficiency and computational cost.

[0118] It should be noted that the route planning of this invention is based on traditional forklift (manual driving) route planning. Factors to consider include: route length, waiting time, priority avoidance, acceleration / deceleration, turning, heavy / empty load, forklift transportation costs, balance of manual operations, and timeliness of warehouse operations. Based on these factors, a feasible dynamic local route planning strategy mechanism needs to be established on the basis of static route planning. This invention establishes the following dynamic route planning strategy based on traditional warehouse operation management: Warehouse management scheduling includes inbound scheduling, outbound scheduling, and transfer scheduling. Based on actual warehouse operations, this invention defines the forklift's state as follows: 0: Idle state; 1: Completed task and heading to the waiting area; 2: Empty and heading to the inbound loading area; 3: Heavy load and heading to the inbound storage location; 4: Empty load and heading to the outbound storage location; 5: Outbound in progress; 6: Heavy load and heading to the unloading area; 7: Empty load and heading to the transfer / outbound storage location; 8: Transfer in progress; 9: Heavy load and heading to the transfer / inbound storage area.

[0119] Based on the above-mentioned inbound, outbound, and transfer operations, each corresponds to a complete task status process:

[0120] The status flow for inbound operations is: 0-2-3-1-0; the status flow for outbound operations is: 0-4-5-6-1-0; and the status flow for warehouse operations is: 0-7-8-9-1-0.

[0121] Considering factors such as orderly warehouse operations, rational allocation of forklift resources, balanced workload for picking personnel, and forklift wear and tear, when a forklift receives a task, it obtains the forklifts with a current status of 0 and their latest status update time, selecting the forklift with the earliest status update time to execute the current task. This invention requires real-time route planning for forklifts receiving tasks and forklifts that need to replan their routes to avoid obstacles.

[0122] S104, calculate the total travel cost Yn of the feasible route. The total travel cost Yn = travel cost Sn + turning cost Zn. The route with the minimum total travel cost Yn is determined as the latest forklift path, and the latest forklift path is added to the warehouse forklift latest path dataset.

[0123] Specifically, for each route in the feasible route dataset, the travel cost Sn is first calculated. The travel cost is directly related to the route length. By accumulating the distances between adjacent coordinate points on the route and combining this with the forklift's energy consumption model (dynamically adjusting the energy consumption coefficient based on forklift type, load, travel speed, etc.), the travel energy cost of the route is obtained and used as the travel cost Sn. Next, the turning cost Zn is calculated. Based on the forklift's mechanical structure characteristics and handling performance, corresponding scores are assigned to turns at different angles (e.g., a 90-degree turn has a higher score than a 45-degree turn, and a sharp turn has an even higher score). The sum of the scores for all turning operations on the route is used as the turning cost Zn. The travel cost Sn and the turning cost Zn for each route are then added together to obtain the total travel cost Yn. By comparing the Yn values ​​of all routes in the feasible route dataset, the route with the smallest Yn value is selected as the latest path for the forklift. This path ensures transportation efficiency while also taking into account the stability of forklift operation and energy economy. It is stored in the warehouse forklift latest path dataset for subsequent real-time navigation and path monitoring.

[0124] S105: Monitor the pairs of paths in the latest path dataset of warehouse forklifts to determine if there is a collision conflict. If so, determine the collision type and calculate the avoidance cost. Based on the collision type and avoidance cost, determine the forklift to avoid and the avoidance strategy corresponding to the forklift to avoid.

[0125] The data twin generation method based on the improved ant colony algorithm monitors collisions on each pair of forklift paths, accurately determines the collision type, and calculates avoidance costs. This allows for the rational arrangement of forklift avoidance strategies, preventing work stoppages and cargo damage caused by forklift collisions, improving the safety and smoothness of warehousing operations, and reducing operational risks.

[0126] The avoidance strategy determines the waiting forklifts, the waiting time, and the spatial location of the waiting area, achieving conflict-free route planning through global avoidance supervision. To better execute the avoidance rules, the avoidance parameters set in this invention are as follows:

[0127] The normal forklift travel speed is preset and converted to X steps / minute to improve calculation efficiency;

[0128] Set the maximum shelf-listing time to T1 minutes, the maximum shelf-delisting time to T2 minutes, the maximum shelf-delisting time for relocation to T3 minutes, and the maximum shelf-delisting time for relocation to T4 minutes;

[0129] For a straight-line collision, the rerouting and avoidance step size is set to E for unloaded vehicles and F for heavily loaded vehicles.

[0130] For different routes that are expected to collide head-on, the unloaded avoidance step size is set to G, and the heavy-load avoidance step size is set to H;

[0131] For perpendicular collisions, the unloaded avoidance step size is set to M, and the heavy-load avoidance step size is set to N;

[0132] For a collision while chasing, the unloaded avoidance step size is set to I, and the heavy-load avoidance step size is set to J;

[0133] The collision types include straight-on collisions, collisions expected to occur on different routes, perpendicular collisions, and chase collisions.

[0134] When the collision type is a straight-on collision:

[0135] If a heavily loaded forklift encounters an unloaded forklift, the route for the unloaded forklift will be replanned at a step E away from the point of collision.

[0136] If the forklifts are under the same load, the forklift with the lower avoidance cost should re-plan its route at a step distance E or F from the collision point.

[0137] If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will re-plan its route at a step length E or F away from the collision point.

[0138] When the collision type is a predicted head-on collision along different routes:

[0139] If a heavily loaded forklift encounters an unloaded forklift, the unloaded forklift will wait to avoid the collision at a distance G from the point of impact.

[0140] If the forklifts are under the same load, calculate the collision location, and the forklift with the lower avoidance cost will wait to avoid the collision at a step length G or H away from the collision point.

[0141] If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will wait to avoid the collision at a step length G or H set at the point of collision.

[0142] When the collision type is vertical collision:

[0143] If a heavily loaded forklift encounters an unloaded forklift, the unloaded forklift will wait to avoid the collision at a distance M from the point of impact.

[0144] If the forklifts are under the same load, calculate the collision location and have the forklift with the lower avoidance cost wait to avoid the collision at a step distance M or N from the collision point.

[0145] If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will wait to avoid the collision at a step distance M or N from the collision point.

[0146] When the collision type is chasing collision:

[0147] When the following vehicle yields, it maintains a preset step length of I or J. After the preceding vehicle turns and moves away from the preset distance, it continues to drive normally.

[0148] The data twin generation method based on the improved ant colony algorithm meticulously categorizes collision types into straight-line head-on collisions, predicted head-on collisions along different routes, perpendicular collisions, and chasing collisions. For each type, differentiated avoidance strategies are developed based on factors such as forklift load status, avoidance costs, and task initiation time. This makes collision handling more scientific and reasonable, avoiding efficiency losses or safety hazards caused by a one-size-fits-all approach. Through precise avoidance strategies, forklifts can react quickly and orderly when encountering collision risks, maintaining the continuity of warehouse operations, reducing overall operational delays caused by avoidance, ensuring timely entry and exit of goods, and improving the level of warehouse operation services.

[0149] The data twin generation method based on the ant colony algorithm improvement also includes:

[0150] Based on actual warehouse operations, we define 0 as the forklift idle state, 1 as the forklift completing its task and heading to the waiting area, 2 as the forklift empty and heading to the inbound loading area, 3 as the forklift heavily loaded and heading to the inbound storage location, 4 as the forklift empty and heading to the outbound storage location, 5 as the forklift in the outbound process, 6 as the forklift heavily loaded and heading to the unloading area, 7 as the forklift empty and heading to the transfer outbound storage location, 8 as the forklift in the transfer process, and 9 as the forklift heavily loaded and heading to the transfer inbound storage location.

[0151] Define 00 as goods awaiting warehousing, 11 as goods in warehousing and transportation status, 22 as goods already in warehousing status, 33 as goods awaiting transfer to warehousing status, 44 as goods transferred and removed from shelves status, 55 as goods in warehousing and transportation status, 66 as other goods transferred and removed from shelves status, 77 as goods awaiting shipment status, 88 as goods shipped and removed from shelves status, 99 as goods shipped and transportation status, 100 as goods already shipped status, and 110 as other goods shipped and removed from shelves status.

[0152] The status flow for forklift inbound operations is 0-2-3-1-0:

[0153] Receive the inbound business order order1 (O1, O2, ... On), associate the goods An with the business order On, and set the status of goods An to 00;

[0154] Get the forklift with a status of 0, assign a forklift Cn, and associate the business order On with the forklift Cn;

[0155] Based on all the data of warehouse locations that are not full, a recommended location Kx is obtained. Based on the recommended location Kx, the starting point and target point of the forklift to perform the task are determined. Based on the starting point and target point, a route is planned for the forklift.

[0156] Forklift Cn goes to the pickup area to scan the tag code of goods An and associates it with goods An. The status of goods An is 11.

[0157] After the forklift Cn transports the goods to the warehouse location, the label code of goods An is scanned a second time to confirm that goods An has been put on the shelf.

[0158] The business scheduling platform receives the scanned information and confirms the actual storage location Kn where the forklift is placed.

[0159] If the business scheduling platform obtains the number N of goods in storage location Kn with a status of 22, then N+1 is the layer Hn of goods An in storage location Kn with a status of 22.

[0160] The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to create a digital twin model of the goods.

[0161] The system automatically associates orders, forklifts, and goods, planning pickup, transportation, and shelving routes. Combined with barcode scanning to confirm the location and status of goods, it achieves a high degree of automation and collaboration in the goods receiving process, reducing manual intervention, lowering the probability of errors, and improving receiving efficiency. Simultaneously, it generates real-time digital twin models of goods, facilitating visualized warehouse management. Based on data indicating insufficient warehouse locations, it recommends storage locations and, combined with forklift barcode scanning and information interaction with the business scheduling platform, accurately records the location and layer of goods, providing precise inventory information for warehouse management. This facilitates goods retrieval and inventory counting, optimizes warehouse space utilization, and improves inventory management accuracy.

[0162] The outbound status flow for forklifts is 0-4-5-6-1-0:

[0163] Receive the transfer order order2(O1, O2, ... On), associate the goods An with the order On, and set the status of goods An to 33;

[0164] Get the forklift with a status of 0, assign a forklift Cn, and associate the current order On with the forklift Cn;

[0165] Obtain the current storage location K of goods An and the target storage location K1 for relocation. Based on the target storage location K1 and the current storage location K, determine the starting point and target point of the forklift to perform the task, and plan the route for the forklift based on the starting point and target point.

[0166] The business scheduling platform obtains the layer number Fn of goods An in the current storage location K and the total layer number F of goods in the current storage location. F-Fn represents the goods (Ay1, Ay2, ...) above goods An.

[0167] Determine if the number of layers Fn is equal to the total number of layers F. If so, the forklift scans the barcode to remove goods An from the shelf. The goods status of goods An is 55. The forklift moves goods An to the target storage location K1, puts it on the shelf, and scans the barcode of goods An. The business scheduling platform receives the barcode information and confirms the actual storage location Kn where the forklift puts it on. The business scheduling platform obtains the number N of goods in storage location Kn with a status of 22. Then N+1 is the number of layers Hn of goods An in storage location Kn with a status of 22. The data platform generates the storage location and corresponding layer number of each goods in the warehouse in real time and generates a digital twin model of the goods.

[0168] If not, the forklift scans the barcode and gradually removes the goods (Ay1, Ay2...) above Goods An. The status of the goods (Ay1, Ay2...) above Goods An is 66. The forklift scans the barcode and removes Goods An. The status of Goods An is 44. The forklift scans the barcode and gradually puts the goods (Ay1, Ay2...) above Goods An back on. The layer number of the goods (Ay1, Ay2...) above Goods An is the original layer number minus 1, and the status is 22. The forklift scans the barcode... Goods An has a status of 55. A forklift moves to the target storage location K1, puts the goods An on the shelf, and scans the barcode. The business scheduling platform receives the barcode information and confirms the actual storage location Kn where the forklift puts the goods on the shelf. The business scheduling platform obtains the number N of goods in storage location Kn with a status of 22. Then N+1 is the layer Hn of goods An in storage location Kn. The status of goods An is 22. The data platform generates the storage location and corresponding layer number of each goods in the warehouse in real time, generating a digital twin model of the goods.

[0169] For goods transfer operations, the system also relies on status codes and the association mechanism between orders, forklifts, and goods. Combined with changes in storage location, it precisely plans transfer routes and intelligently handles the order of goods being put on and taken off shelves based on their layer position in the original storage location. This ensures efficient and accurate transfer operations, reduces the number of times goods are handled, lowers the risk of damage, and updates the digital twin model of goods in real time to reflect inventory dynamics. Whether transferring goods to the top or middle layers, the system can adapt to different storage layouts and goods stacking conditions by rationally arranging the scanning and putting-off steps. It flexibly handles complex scenarios in the transfer process, ensuring smooth transfer operations and maintaining the continuity of warehouse operations.

[0170] The forklift transfer process status is 0-7-8-9-1-0:

[0171] Receive a request for goods An to be shipped out, associate the goods An with the transfer order order3 (O1, O2, ... On), associate the current order On with the current goods An, and the goods status is 77;

[0172] Get the forklift with a status of 0, assign a forklift Cn, and associate the current order On with the forklift Cn;

[0173] Obtain the current storage location K of goods An, determine the starting point and target point of the forklift to perform the task based on the current storage location K, and plan the route for the forklift based on the starting point and target point;

[0174] The business scheduling platform obtains the layer number Fn of goods An in storage location K and the total layer number F of goods in the current storage location, where F-Fn is the goods (Ay1, Ay2, ...) above goods An.

[0175] Determine if the number of layers Fn is equal to the total number of layers F. If so, the forklift scans the barcode of goods An and removes them from the shelf, setting the status of goods An to 99. The forklift then proceeds to the outbound area, completes the outbound process, and scans the barcode of goods An. The business scheduling platform receives the barcode information and confirms that the forklift is in the outbound picking area. The business scheduling platform marks the status of goods An as 100. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to generate a digital twin model of the goods.

[0176] If not, the forklift scans the goods (Ay1, Ay2...) above goods An and gradually removes them from the shelves, making the goods status of goods (Ay1, Ay2...) above goods An 110. The forklift scans the goods An to remove them from the shelves, making the goods status of goods An 88. The forklift scans the goods (Ay1, Ay2...) above goods An and gradually puts them back on the shelves, making the goods status of goods (Ay1, Ay2...) above goods An 22. The layer number is the original record layer number minus 1, and the status is 22. The forklift scans goods An, and the goods status of goods An is 99. The forklift moves to the outbound area to complete the outbound process and scans goods An. The business scheduling platform receives the scanning information and confirms that the forklift is in the outbound picking area. The business scheduling platform marks the status of goods An as 100. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to generate a digital twin model of the goods.

[0177] The system systematically links orders, forklifts, and goods, plans precise routes based on goods storage locations, and intelligently processes unloading operations according to the number of product layers. This ensures goods are quickly and accurately unloaded and transported to the outbound area. Outbound status is confirmed by scanning barcodes, automating and maximizing the outbound process. The business scheduling platform and data platform collaborate throughout the process, collecting and updating goods storage location, layer, and status information in real time. This generates a digital twin model of the goods, allowing warehouse managers to monitor outbound dynamics accurately and in real time, facilitating timely adjustments to warehousing strategies and improving the level of visualized management in warehouse operations.

[0178] Figure 2 This is a schematic diagram illustrating the main components of the data twin generation system based on the ant colony algorithm improved in this embodiment of the invention. Figure 2As shown, the data twin generation system 1 based on the improved ant colony algorithm provided in this embodiment of the invention includes a forklift task allocation module 10, a path planning request and response module 20, a feasible route generation module 30, an optimal path determination module 40, and a collision monitoring and avoidance module 50.

[0179] The forklift task allocation module 10 is used to respond to task requests, obtain the status information of all forklifts, select the forklift whose status information is idle and whose update time is the earliest to execute the task, and associate the forklift executing the task with the current task order. The tasks include goods inbound tasks, goods transfer tasks, and goods outbound tasks.

[0180] The path planning request and response module 20 is used to respond to the route planning request of the forklift performing the task and determine the starting point and target point of the forklift performing the task.

[0181] The feasible route generation module 30 is used to generate multiple feasible routes based on the relative positions of the starting point and the target point, and put the multiple feasible routes into the feasible route dataset;

[0182] The optimal route determination module 40 is used to calculate the total travel cost Yn of feasible routes, determine the route with the minimum total travel cost Yn as the latest route of the forklift, and put the latest route into the warehouse forklift latest route dataset.

[0183] The collision detection and avoidance module 50 is used to monitor the pairs of paths in the latest path data of warehouse forklifts, determine whether a collision conflict has occurred, determine the collision type if so, calculate the avoidance cost, obtain the avoidance forklift status based on the dynamic priority avoidance strategy, determine whether the forklift needs to replan its route, and if so, dynamically plan the minimum operating cost path for the forklift until the minimum cost path planning for all forklifts is completed.

[0184] Figure 8 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as... Figure 8 As shown, the electronic device 60 includes: a processor 601, a memory 602, and a bus 603;

[0185] The processor 601 and the memory 602 communicate with each other via the bus 603.

[0186] The processor 601 is used to call program instructions in the memory 602 to execute the methods provided in the above-described method embodiments, and to execute the methods provided in the embodiments of the present invention.

[0187] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions, which cause the computer to execute the methods provided in the above-described method embodiments.

[0188] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various storage media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0189] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A data twin generation method based on an improved ant colony algorithm, characterized in that, include: In response to a task request, the status information of all forklifts is obtained, and the forklift with the idle status information and the earliest update time is selected to execute the task. The forklift executing the task is associated with the current task, wherein the task includes goods inbound task, goods transfer task, and goods outbound task. In response to a forklift route planning request for a task, determine the starting point and destination point of the forklift. Based on the relative positions of the starting point and the target point, multiple feasible routes are generated and placed into a feasible route dataset. The total driving cost Yn of a feasible route is calculated by accumulating the distances between adjacent coordinate points on the route and combining this with the forklift's energy consumption model to obtain the driving energy cost of the route, which is then used as the driving cost Sn. Based on the mechanical structure characteristics and handling performance of the forklift, corresponding scores are assigned to turns at different angles, and the sum of the scores of all turning operations on the route is used as the turning cost Zn. The total driving cost Yn = driving cost Sn + turning cost Zn. The route with the minimum total driving cost Yn is determined as the latest forklift path, and the latest forklift path is added to the warehouse forklift latest path dataset. Monitor the pairs of paths in the latest path dataset of warehouse forklifts to determine if there is a collision conflict. If so, determine the collision type and calculate the avoidance cost. Based on the collision type and avoidance cost, determine the forklift to avoid and the avoidance strategy corresponding to the forklift to avoid.

2. The data twin generation method based on the improved ant colony algorithm according to claim 1, characterized in that, The data twin generation method based on the ant colony algorithm improvement also includes: Project the warehouse digital raster map onto the raster matrix so that the forklift's movement step is equal to the side length of the raster. Mark the grid coordinates corresponding to the lanes as feasible areas, mark the grid coordinates corresponding to the parking spaces as infeasible areas, and mark the grid coordinates between the lanes and parking spaces as infeasible areas to construct a grid matrix. Initialize the pheromone matrix Tau, and set the maximum number of iterations K, the number of ants M, the pheromone importance parameter, the heuristic factor importance parameter, the pheromone evaporation coefficient, and the pheromone increase intensity coefficient; Establish a matrix D to store the cost of each grid point to its adjacent unobstructed grid points; Construct a heuristic information matrix, using the reciprocal of the distance from the current grid point to the target grid point as the heuristic information, and setting the heuristic information at obstacles to 0.

3. The data twin generation method based on the ant colony algorithm improvement according to claim 2, characterized in that, Based on the relative positions of the starting point and the target point, multiple feasible routes are generated, and these routes are added to a feasible route dataset, including: For each ant, initialize the ant's crawling path and path length. The ant starts from the starting point and begins to search for the path, finding all unobstructed adjacent grid points of the current grid point, and filtering the available grid points based on the taboo list. If the current grid point is not the target grid point and the number of selectable grid points is greater than or equal to 1, calculate the probability that the ant will move from the current grid point to each of the adjacent grid points, select the next grid point based on the roulette wheel method, update the ant's crawling path, path length, matrix D and taboo list, until the current grid point is the target grid point or the number of selectable grid points is less than 1, end the current iteration, and record the ant's walking route; If the ant's last step is the target grid point, then calculate the current path length and compare it with the currently known shortest path length. If the current path length is less than the currently known shortest path length, then update the currently known shortest path length and the shortest path. If the ant's last step is not the target grid point, then the path length is recorded as 0; Repeat the path search steps until M ants have completed one round of path search. Determine if the termination condition is met. If so, end the ant colony algorithm and draw the optimal planned path.

4. The data twin generation method based on the improved ant colony algorithm according to claim 1, characterized in that, The collision types include straight-on collisions, collisions expected to occur on different routes, perpendicular collisions, and chase collisions. When the collision type is a straight-on collision: If a heavily loaded forklift encounters an unloaded forklift, the route for the unloaded forklift will be replanned at a step E away from the point of collision. If the forklifts are under the same load, the forklift with the lower avoidance cost should re-plan its route at a step distance E or F from the collision point. If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will re-plan its route at a step length E or F away from the collision point. When the collision type is a predicted head-on collision along different routes: If a heavily loaded forklift encounters an unloaded forklift, the unloaded forklift will wait to avoid the collision at a distance G from the point of impact. If the forklifts are under the same load, calculate the collision location, and the forklift with the lower avoidance cost will wait to avoid the collision at a step length G or H away from the collision point. If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will wait to avoid the collision at a step length G or H set at the point of collision. When the collision type is vertical collision: If a heavily loaded forklift encounters an unloaded forklift, the unloaded forklift will wait to avoid the collision at a distance M from the point of impact. If the forklifts are under the same load, calculate the collision location and have the forklift with the lower avoidance cost wait to avoid the collision at a step distance M or N from the collision point. If the forklifts have the same load and the same avoidance cost, the forklift that initiates the task later will wait to avoid the collision at a step distance M or N from the collision point. When the collision type is chasing collision: When the following vehicle yields, it maintains a preset step length of I or J. After the preceding vehicle turns and moves away from the preset distance, it continues to drive normally.

5. The data twin generation method based on the ant colony algorithm improvement according to claim 1, characterized in that, The data twin generation method based on the ant colony algorithm improvement also includes: Based on actual warehouse operations, we define 0 as the forklift idle state, 1 as the forklift completing its task and heading to the waiting area, 2 as the forklift empty and heading to the inbound loading area, 3 as the forklift heavily loaded and heading to the inbound storage location, 4 as the forklift empty and heading to the outbound storage location, 5 as the forklift in the outbound process, 6 as the forklift heavily loaded and heading to the unloading area, 7 as the forklift empty and heading to the transfer outbound storage location, 8 as the forklift in the transfer process, and 9 as the forklift heavily loaded and heading to the transfer inbound storage location. Define 00 as goods awaiting warehousing, 11 as goods in warehousing and transportation status, 22 as goods already in warehousing status, 33 as goods awaiting transfer to warehousing status, 44 as goods transferred and removed from shelves status, 55 as goods in warehousing and transportation status, 66 as other goods transferred and removed from shelves status, 77 as goods awaiting shipment status, 88 as goods shipped and removed from shelves status, 99 as goods shipped and transportation status, 100 as goods already shipped status, and 110 as other goods shipped and removed from shelves status. The status flow for forklift inbound operations is 0-2-3-1-0, the status flow for forklift outbound operations is 0-4-5-6-1-0, and the status flow for forklift transfer operations is 0-7-8-9-1-0.

6. The data twin generation method based on the ant colony algorithm improvement according to claim 1, characterized in that, The data twin generation method based on the ant colony algorithm improvement also includes: Receive the inbound business order order1 (O1, O2, ... On), associate the goods An with the business order On, and set the status of goods An to 00; Get the forklift with a status of 0, assign a forklift Cn, and associate the business order On with the forklift Cn; Based on all the data of warehouse locations that are not full, a recommended location Kx is obtained. Based on the recommended location Kx, the starting point and target point of the forklift to perform the task are determined. Based on the starting point and target point, a route is planned for the forklift. Forklift Cn goes to the pickup area to scan the tag code of goods An and associates it with goods An. The status of goods An is 11. After the forklift Cn transports the goods to the warehouse location, the label code of goods An is scanned a second time to confirm that goods An has been put on the shelf. The business scheduling platform receives the scanned information and confirms the actual storage location Kn where the forklift is placed. The business scheduling platform obtains the number N of goods N in storage location Kn with a status of 22. Then N + 1 is the layer Hn of goods An in storage location Kn with a status of 22. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to create a digital twin model of the goods.

7. The data twin generation method based on the ant colony algorithm improvement according to claim 5, characterized in that, The data twin generation method based on the ant colony algorithm improvement also includes: Receive the transfer order order2 (O1, O2, ... On), associate the goods An with the order On, and set the status of goods An to 33; Get the forklift with a status of 0, assign a forklift Cn, and associate the current order On with the forklift Cn; Obtain the current storage location K of goods An and the target storage location K1 for relocation. Based on the target storage location K1 and the current storage location K, determine the starting point and target point of the forklift to perform the task, and plan the route for the forklift based on the starting point and target point. The business scheduling platform obtains the layer number Fn of goods An in the current storage location K and the total layer number F of goods in the current storage location. F-Fn represents the goods (Ay1, Ay2, ...) above goods An. Determine if the number of layers Fn is equal to the total number of layers F. If so, the forklift scans the barcode to remove goods An from the shelf. The goods status of goods An is 55. The forklift moves goods An to the target storage location K1, puts it on the shelf, and scans the barcode of goods An. The business scheduling platform receives the barcode information and confirms the actual storage location Kn where the forklift puts it on. The business scheduling platform obtains the number N of goods in storage location Kn with a status of 22. Then N + 1 is the number of layers Hn of goods An in storage location Kn with a status of 22. The data platform generates the storage location and corresponding layer number of each goods in the warehouse in real time, generating a digital twin model of the goods. If not, the forklift scans the barcode and gradually removes the goods (Ay1, Ay2...) above goods An. The status of the goods (Ay1, Ay2...) above goods An is 66. The forklift scans the barcode and removes goods An. The status of goods An is 44. The forklift scans the barcode and gradually puts the goods (Ay1, Ay2...) above goods An back on the shelf. The layer number of the goods (Ay1, Ay2...) above goods An is the original record layer number - 1, and the status is 22. The forklift scans the barcode of goods An. The status of goods An is 55. The forklift moves to the target storage location K1, puts on the shelf and scans the barcode of goods An. The business scheduling platform receives the scanning information and confirms the actual storage location Kn where the forklift puts the shelf. The business scheduling platform obtains the number N of goods in storage location Kn with a status of 22. Then N + 1 represents the layer number Hn of goods An in storage location Kn. The status of goods An is 22. The data platform generates the storage location and corresponding layer number of each goods in the warehouse in real time, generating a digital twin model of the goods.

8. The data twin generation method based on the ant colony algorithm improvement according to claim 6, characterized in that, The data twin generation method based on the ant colony algorithm improvement also includes: Upon receiving a request for goods An to be shipped out, the goods An are associated with a transfer order order3 (O1, O2, ... On). The current order On is associated with the current goods An, and the goods status is 77. Get the forklift with a status of 0, assign a forklift Cn, and associate the current order On with the forklift Cn; Obtain the current storage location K of goods An, determine the starting point and target point of the forklift to perform the task based on the current storage location K, and plan the route for the forklift based on the starting point and target point; The business scheduling platform obtains the layer number Fn of goods An in storage location K and the total layer number F of goods in the current storage location. F-Fn represents the goods (Ay1, Ay2, ...) above goods An. Determine if the number of layers Fn is equal to the total number of layers F. If so, the forklift scans the barcode of goods An and removes them from the shelf, setting the status of goods An to 99. The forklift then proceeds to the outbound area, completes the outbound process, and scans the barcode of goods An. The business scheduling platform receives the barcode information and confirms that the forklift is in the outbound picking area. The business scheduling platform marks the status of goods An as 100. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to generate a digital twin model of the goods. If not, the forklift scans the goods (Ay1, Ay2...) above goods An and gradually removes them from the shelves, making the goods status of goods (Ay1, Ay2...) above goods An 110. The forklift scans the goods An to remove them from the shelves, making the goods status of goods An 88. The forklift scans the goods (Ay1, Ay2...) above goods An and gradually puts them back on the shelves, making the goods status of goods (Ay1, Ay2...) above goods An 22. The layer number is the original record layer number minus 1, and the status is 22. The forklift scans goods An, and the goods status of goods An is 99. The forklift moves to the outbound area to complete the outbound process and scans goods An. The business scheduling platform receives the scanning information and confirms that the forklift is in the outbound picking area. The business scheduling platform marks the status of goods An as 100. The data platform generates the storage location and corresponding layer number of each item in the warehouse in real time to generate a digital twin model of the goods.

9. A system for generating data twins based on an improved ant colony algorithm, characterized in that, include: The forklift task allocation module is used to respond to task requests, obtain the status information of all forklifts, select the forklift whose status information is idle and whose update time is the earliest to execute the task, and associate the forklift executing the task with the current task order. The tasks include goods receiving tasks, goods transfer tasks, and goods outgoing tasks. The path planning request and response module is used to respond to the forklift route planning request for performing the task and determine the starting point and target point of the forklift performing the task. The feasible route generation module is used to generate multiple feasible routes based on the relative positions of the starting point and the target point, and put the multiple feasible routes into the feasible route dataset; The optimal path determination module is used to calculate the total travel cost Yn of feasible routes. By accumulating the distances between adjacent coordinate points on the route and combining the forklift's energy consumption model, the travel energy cost of the route is obtained, which is used as the travel cost Sn. Based on the mechanical structure characteristics and handling performance of the forklift, corresponding scores are assigned to turns at different angles, and the sum of the scores of all turning operations on the route is used as the turning cost Zn. The total travel cost Yn = travel cost Sn + turning cost Zn. The route with the minimum total travel cost Yn is determined as the latest path of the forklift, and the latest path is added to the warehouse forklift latest path dataset. The collision detection and avoidance module is used to monitor the pairs of paths in the latest path dataset of warehouse forklifts, determine whether a collision has occurred, and if so, determine the collision type and calculate the avoidance cost. Based on the dynamic priority avoidance strategy, it obtains the status of the forklifts to avoid the collision and determines whether the forklifts need to replan their routes. If so, it dynamically plans the path with the minimum operating cost for the forklifts until the minimum cost path planning for all forklifts is completed.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

Images