An AGV-based intelligent material distribution management optimization method and system

Abstract

The present invention discloses an intelligent material distribution management optimization method and system based on AGV, which relates to the field of AGV scheduling management technology, including: acquiring and analyzing factory operation data to obtain material demand forecast results; the factory operation data includes material demand data and AGV status data; based on the factory operation data, using a multi-objective optimization algorithm to generate a first AGV distribution task sequence; based on the material demand forecast results and the first AGV distribution task sequence, generating a second AGV distribution task sequence, controlling the AGV to perform material distribution tasks according to the second AGV distribution task sequence, and simultaneously monitoring task execution. The present invention realizes intelligent optimization of material distribution management by combining an LSTM network for material demand forecasting, adopting a multi-objective optimization algorithm for task sequence generation, and real-time monitoring of AGV execution status, significantly improving the operating efficiency of the AGV distribution system, reducing energy consumption, and achieving dynamic balancing of equipment loads.

Description

Technical Field

[0001] The present invention relates to the technical field of AGV scheduling management, and in particular to an AGV-based intelligent material distribution management optimization method and system. Background Art

[0002] With the deepening development of intelligent manufacturing, the Industry 4.0 era has placed higher demands on material distribution management. As the core equipment of smart factory material distribution systems, the optimization of AGV (Automated Guided Vehicle) scheduling and management systems directly impacts production efficiency and operating costs. Traditional AGV scheduling management primarily utilizes rule-based heuristic algorithms or simple real-time response strategies. These approaches have significant limitations when dealing with complex and changing material demands. In recent years, with the advancement of artificial intelligence technology, some researchers have begun to explore the application of machine learning algorithms to AGV scheduling optimization, such as using reinforcement learning for path planning and neural networks for material demand prediction. However, these methods often separate material demand prediction, task planning, and real-time scheduling, lacking a systematic, holistic optimization solution.

[0003] At present, the AGV material distribution management system has the following main technical problems: First, most of the existing material demand forecasting methods use traditional time series analysis or simple statistical models, which fail to fully consider the cyclical characteristics and long-term dependencies of material demand, resulting in insufficient prediction accuracy; Second, in the task planning stage, existing algorithms often only consider a single optimization goal (such as distribution efficiency), while ignoring multi-dimensional indicators such as energy consumption and load balancing, making it difficult to achieve global optimization; Third, in the real-time scheduling process, due to the lack of effective utilization of the real-time status of AGVs and dynamic optimization capabilities, the system has poor adaptability when facing emergencies, which can easily cause delivery delays or waste of resources; Finally, the existing system generally has the phenomenon of information islands, and the coordination between various functional modules is insufficient, making it difficult to achieve end-to-end intelligent optimization.

[0004] Therefore, there is an urgent need for an intelligent material distribution management optimization method that can effectively solve the above-mentioned problems and improve the overall efficiency of the AGV distribution system. The AGV-based intelligent material distribution management optimization method and system proposed in this paper addresses the shortcomings of existing technologies and innovates in multiple dimensions, including material demand forecasting, multi-objective task optimization, and real-time monitoring. It belongs to the field of intelligent manufacturing and AGV scheduling management technology. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the present invention provides an AGV-based intelligent material distribution management optimization method and system, which can solve the problems mentioned in the background technology.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: an AGV-based intelligent material distribution management optimization method, comprising: acquiring and analyzing factory operation data to obtain material demand forecast results; the factory operation data includes material demand data and AGV status data; based on the factory operation data, a first AGV distribution task sequence is generated using a multi-objective optimization algorithm; based on the material demand forecast results and the first AGV distribution task sequence, a second AGV distribution task sequence is generated, and the AGV is controlled to perform the material distribution task according to the second AGV distribution task sequence, and task execution monitoring is performed at the same time.

[0008] As a preferred solution of the AGV-based intelligent material distribution management optimization method described in the present invention, the process of obtaining the material demand forecast result includes the following steps: classifying the material demand data according to the material type to obtain material classification data; extracting the material demand cycle characteristics in the material classification data; based on the material demand cycle characteristics, using the long short-term memory network LSTM for training to obtain a material demand forecast model; calculating the AGV distribution capacity data based on the AGV status data; generating the material demand forecast result based on the material demand forecast model, material cycle characteristics and AGV distribution capacity data.

[0009] As a preferred solution of the AGV-based intelligent material distribution management optimization method described in the present invention, the material demand data includes material type, quantity and urgency; the AGV status data includes current location, power and load conditions; the material demand cycle characteristics extraction process is specifically to analyze the material classification data through the autocorrelation function, find the main cycle, and construct a cycle characteristic function, and obtain the material cycle characteristics based on the cycle characteristic function.

[0010] As a preferred solution of the AGV-based intelligent material distribution management optimization method described in the present invention, wherein: based on factory operation data, a first AGV distribution task sequence is generated using a multi-objective optimization algorithm, including the following steps: based on the factory operation data, a set of objective functions is constructed; based on the objective function set, the conflict degree between AGV distribution tasks is determined, and the task priority is determined according to the conflict degree; based on the objective function set and the task priority, a multi-population co-evolution algorithm is executed; based on the global non-dominated solution set and the task priority, a comprehensive score of each solution is calculated; based on the comprehensive score, the optimal solution is selected to generate the first AGV distribution task sequence, and the first AGV distribution task sequence includes AGV number, task number and execution time information.

[0011] As a preferred solution of the AGV-based intelligent material distribution management optimization method described in the present invention, the objective function set includes a distribution efficiency objective function, an energy consumption objective function, an equipment load balancing objective function and an urgency objective function.

[0012] The conflict degree is calculated comprehensively by time difference, spatial distance and processing time difference, and the formula is expressed as follows:

[0013] ;

[0014] Where, 、 are the expected start time of the i-th delivery task and the j-th delivery task, respectively, where the i-th delivery task and the j-th delivery task constitute a task pair; The maximum expected start time difference between all delivery task pairs; is the spatial distance between the i-th delivery task and the j-th delivery task; is the maximum spatial distance between all delivery task pairs; 、 are the estimated processing time of the i-th delivery task and the j-th delivery task respectively; is the maximum processing time difference between all delivery task pairs; 、 、 are the weight coefficients of the three dimensions of time difference, spatial distance and processing time.

[0015] The task priority is determined by the ratio of task urgency to conflict degree:

[0016] ;

[0017] Where, is the priority value of the i-th delivery task; is the sum of the conflict degrees between the ith delivery task and all other delivery tasks; u i The urgency of the task.

[0018] As a preferred solution of the AGV-based intelligent material distribution management optimization method of the present invention, the multi-population co-evolution algorithm includes:

[0019] Construct K subpopulations:

[0020] ;

[0021] in, is the number of subpopulations;

[0022] For each of the subpopulations , when its optimal solution has not been updated in N consecutive generations, calculate the population diversity index:

[0023] ;

[0024] in, is the size of the i-th subpopulation; For subpopulation individuals in is the number of objective functions; is the evaluation value of the j-th objective function on individual x; is the jth objective function in the population The average value in is the diversity index value of the i-th subpopulation;

[0025] like , then perform cross-population recombination:

[0026] ;

[0027] in, are individuals from the current population, are individuals from the non-dominated solution set of other populations; It is a new individual produced after recombination; is the reorganization weight coefficient; is the threshold of the diversity indicator.

[0028] like , then perform adaptive mutation:

[0029] ;

[0030] in, , g is the current algebra, is the maximum algebra, b is the shape parameter; is the new individual after mutation; x is the individual before mutation; is the upper bound of the decision variable; is the lower bound of the decision variable; is a random number, ∈[0,1]; Adaptive variable step length.

[0031] Information exchange is performed between populations every M generations, and the global non-dominated solution set is updated by calculating the dominance relationship of the solutions.

[0032] As a preferred solution of the AGV-based intelligent material distribution management optimization method described in the present invention, the monitoring includes AGV position tracking, task completion evaluation and abnormal situation reporting.

[0033] In order to further solve the above technical problems, the present invention provides the following technical solutions: an AGV-based intelligent material distribution management optimization system, comprising: a data analysis module, used to acquire and analyze factory operation data to obtain material demand forecast results; a task generation module, which generates a first AGV distribution task sequence based on the factory operation data using a multi-objective optimization algorithm; a task optimization execution module, which generates a second AGV distribution task sequence based on the material demand forecast results and the first AGV distribution task sequence, controls the AGV to perform material distribution tasks according to the second AGV distribution task sequence, and monitors the task execution at the same time.

[0034] A computer device includes a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, the steps of the above-mentioned AGV-based intelligent material distribution management optimization method are implemented.

[0035] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of the above-mentioned AGV-based intelligent material distribution management optimization method are implemented.

[0036] Beneficial effects of the present invention: The present invention realizes intelligent optimization of material distribution management by combining LSTM network to predict material demand, adopting multi-objective optimization algorithm to generate task sequence and real-time monitoring of AGV execution status, significantly improving the operating efficiency of AGV distribution system, reducing energy consumption and achieving dynamic balancing of equipment load. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 This is an overall flow chart of an AGV-based intelligent material distribution management optimization method provided by one embodiment of the present invention.

[0039] Figure 2 This is a computer equipment diagram of the AGV-based intelligent material distribution management optimization method of the present invention. DETAILED DESCRIPTION

[0040] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0041] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0042] Example 1, reference Figure 1 , as an embodiment of the present invention, provides an AGV-based intelligent material distribution management optimization method.

[0043] Figure 1 The figure shows an overall flow chart of an AGV-based intelligent material distribution management optimization method, including:

[0044] S1: Obtain and analyze factory operation data to obtain material demand forecast results;

[0045] S2: Based on the factory operation data, a multi-objective optimization algorithm is used to generate a first AGV delivery task sequence;

[0046] S3: Based on the material demand forecast result and the first AGV delivery task sequence, a second AGV delivery task sequence is generated, and the AGV is controlled to execute the material delivery task according to the second AGV delivery task sequence, while task execution monitoring is performed.

[0047] This example will describe S1 to S3 in detail one by one:

[0048] S1: Obtain and analyze factory operation data to obtain material demand forecast results.

[0049] Specifically, factory operation data includes material demand data and AGV status data. Material demand data includes material type, quantity, and urgency. AGV status data includes current location, power level, and load.

[0050] It should be noted that, in the present invention, factory operation data is acquired using a layered data acquisition architecture: material demand data is collected in real time through intelligent data acquisition terminals and RFID technology deployed at each workstation; the AGV's onboard positioning system (such as laser SLAM or UWB positioning), battery management system (BMS), and weight sensor are used to respectively collect the AGV's current position, power level, and load status; an industrial-grade wireless network (such as industrial Wi-Fi or 5G private network) and the OPC UA protocol are used to build a reliable data transmission channel, and an on-site data cache server is set up to ensure the reliability of data transmission. Furthermore, it is recognized by those skilled in the art that the collected raw data needs to be subjected to real-time denoising, outlier detection, and standardization processing, and a data quality assessment mechanism needs to be established to provide high-quality data support for subsequent data analysis and prediction models.

[0051] S101: Classify the material demand data according to the material type to obtain material classification data:

[0052] ;

[0053] ;

[0054] Where M is the set of material types; n is the total number of material types; is the material classification data, specifically the global historical demand data of the i-th material; T hist is the time span of global historical demand data; ~ One of the two can be expressed as , which means the demand for the i-th type of material at time point t;

[0055] It should be noted that the classification of material demand data by material type is primarily based on the following classification rules: First, classification is based on the material's physical properties, such as classifying liquid materials, solid bulk materials, and individual items as different types (m1, m2, m3, etc.), as this directly relates to the different types of carriers required for AGVs. Second, classification is based on material storage conditions, such as ambient temperature materials, constant temperature materials, and hazardous chemicals, as this determines the specific conditions that AGVs must meet during distribution. Third, classification is based on the material's use in the production process, such as production raw materials, packaging auxiliary materials, and equipment maintenance materials. This classification reflects the varying timeliness requirements of material demand. Through this multi-dimensional classification approach, the present invention can accurately record and analyze data information for each material type at various points in time, providing more targeted data support for subsequent demand forecasting and AGV mission planning.

[0056] S102: Based on the material classification data, the time series analysis method is used to extract the material demand cycle characteristics.

[0057] Specifically, from the material classification data Extract the periodic characteristics of each type of material. First, use the autocorrelation function Analyze the global historical demand data of material demand and find the main cycles , and construct a periodic characteristic function, and obtain the material periodic characteristics based on the periodic characteristic function.

[0058] ;

[0059] ;

[0060] ;

[0061] in, is the autocorrelation function of the i-th type of material; k is the time lag value; is the average value of the demand for the i-th type of materials; is the main cycle of the i-th type of material; K is the maximum detection cycle length; The material periodic characteristics output by the periodic characteristic function; 、 is the parameter of the periodic characteristic function.

[0062] It should be noted that in the prior art, material demand forecasting often ignores the periodic characteristics of different materials and fails to fully consider the periodic fluctuations in demand, which results in a decrease in the accuracy of the forecast when the material demand has obvious periodicity. Therefore, the present invention introduces the autocorrelation function The cyclical characteristics of material demand are extracted and a cyclical characteristic function is constructed to achieve quantitative analysis of cyclical fluctuations, thereby providing accurate cyclical input data for subsequent material demand forecasting models, solving the problem of accurately capturing cyclical changes by relying solely on time series analysis.

[0063] S103: Based on the material demand cycle characteristics, a long short-term memory network (LSTM) is used for training to obtain a material demand forecasting model.

[0064] Specifically, use the material cycle characteristics and short-term historical demand data As input, the LSTM model is used for training to build a material demand forecasting model. The LSTM model is used to capture the long-term dependencies and periodic fluctuation characteristics in material demand and generate the original material demand forecast value at a future time point. , the formula is:

[0065] ;

[0066] in, For the i-th type of material generated by the LSTM model in the future The original demand forecast value of time; For LSTM model based on material cycle characteristics and short-term historical demand data Generated demand forecast function.

[0067] It should be noted that is the short-term historical demand data, which represents the historical demand data of the i-th material in the last T time steps, and serves as the input of the LSTM model; T is the time step length used for the LSTM model input. and global historical demand data The relationship between short-term historical demand data (also known as material classification data) is as follows: short-term historical demand data is a time window extracted from the global historical demand data and used as input for each model forecast. This time window is typically a fragment of data from the most recent several time steps to reflect current trends and changes.

[0068] Assuming global historical demand data The length of time is T hist , represents the complete demand data sequence from the initial time point to the current time; while the short-term historical demand data The time length is T, which is the time window length of the input LSTM model. To extract short-term historical data from the global historical data, you can follow the following method:

[0069] At time point t, the short-term historical demand data is expressed as:

[0070] ;

[0071] Locate the short-term window starting point. From global historical demand data In , find the data from time point tT to t-1.

[0072] Extract time window data. From global historical demand data Extract the data of the most recent T time steps , constituting short-term historical demand data .

[0073] For example, if T hist =100, indicating that the global historical demand data contains 100 time steps, and T=10, indicating that the short-term time window length is 10, then at time point t=100, the short-term historical demand data is:

[0074] .

[0075] Traditional material demand forecasting methods (such as linear regression or simple time series analysis) lack the ability to handle long-term dependencies, making them difficult to adapt to scenarios with complex and fluctuating material demand. However, the LSTM model is capable of capturing long-term dependencies in time series. By combining LSTM with material cycle characteristics, this invention improves the ability to identify long-term dependencies and cyclical changes, addressing the problem that traditional models cannot account for both long-term and short-term demand fluctuations.

[0076] S104: Calculate AGV delivery capability data based on AGV status data:

[0077] ;

[0078] in, is the AGV delivery capacity data at time point t, indicating the overall delivery capacity of AGV; m is the number of AGVs; is the efficiency coefficient of the j-th AGV; is the load capacity of the jth AGV; is the current power of the j-th AGV; The AGV is fully charged; is the distance from the jth AGV to the target position; is the maximum delivery distance.

[0079] It should be noted that the existing AGV delivery capacity evaluation often uses static parameters and fails to consider the changes in real-time status (such as power, load and location), which leads to the delivery capacity not being consistent with the actual situation. , so that the AGV delivery capacity can dynamically reflect the current status of the AGV, ensuring that the predicted value matches the actual AGV delivery capacity. That is, by introducing the real-time status data of the AGV for calculation, the present invention has dynamic adaptability in the delivery capacity evaluation, especially when the delivery tasks are frequent and the demand fluctuates greatly, it can avoid the scheduling problems caused by insufficient delivery capacity and improve the responsiveness of the system.

[0080] S105: Generate material demand forecast results based on the material demand forecast model, material cycle characteristics and AGV delivery capability data.

[0081] Specifically, the original demand forecast value generated by the LSTM model and the AGV delivery capacity data at time point t are combined , generate the final material demand forecast results By balancing factors Control the weight of material demand forecast and delivery capacity to ensure that the generated demand forecast results are consistent with the current actual delivery capacity of AGV.

[0082] ;

[0083] in, For Class I materials in the future Demand forecast value over time; Is the balance factor used to adjust LSTM prediction and The weight between .

[0084] It should be noted that the demand forecast results of existing technologies are usually single-dimensional, that is, they generate material demand based purely on time series models (such as LSTM, ARIMA, etc.). However, traditional models ignore the actual resource limitations of the distribution system (such as AGV distribution capabilities), resulting in the generated demand forecast results possibly exceeding the actual distribution capabilities of the system, which can easily lead to system overload, distribution delays, and even imbalanced resource scheduling. Especially in complex scenarios with large demand fluctuations and limited distribution resources, the forecast results in existing technologies often cannot be directly converted into reasonable distribution plans. In order to solve the above technical problems, the formula proposed in the present invention not only relies on the LSTM prediction results, but also calculates in real time. By introducing a balancing factor, γ, between demand forecasts and actual delivery capacity as a constraint, the present invention dynamically adjusts the weight of these two factors in different situations, ensuring that forecasts both reflect changing material demand trends and meet the actual delivery system's capacity. This allows the system to generate more reasonable and actionable forecasts based on current resource availability, even during peak demand or resource constraints, ensuring that forecasts are aligned with delivery capacity.

[0085] S2: Based on the factory operation data, a multi-objective optimization algorithm is used to generate the first AGV delivery task sequence.

[0086] Specifically, the multi-objective optimization algorithm comprehensively considers multiple delivery indicators, including delivery efficiency, energy consumption, equipment load, and urgency.

[0087] S201: Constructing a set of objective functions based on factory operation data.

[0088] Specifically, the objective function set includes the distribution efficiency objective function, the energy consumption objective function, the equipment load balancing objective function, and the urgency objective function. The objective function set is expressed as:

[0089] ;

[0090] Among them, the distribution efficiency objective function is expressed as:

[0091] ;

[0092] Where, is the distance of the i-th delivery task; The running speed of AGV when performing the i-th delivery task; is the material loading and unloading time of the i-th delivery task; n is the total number of delivery tasks.

[0093] The energy consumption objective function is expressed as:

[0094] ;

[0095] Where, is the load factor; For unit energy consumption.

[0096] The device load balancing objective function is expressed as:

[0097] ;

[0098] Where, is the number of delivery tasks for the i-th AGV.

[0099] The urgency objective function is expressed as:

[0100] ;

[0101] Where, is the urgency coefficient of the delivery task; is the response time.

[0102] S202: Determine the degree of conflict between AGV delivery tasks based on the objective function set, and determine the task priority based on the conflict degree.

[0103] Specifically, the conflict degree is obtained by comprehensively calculating the time difference, spatial distance and processing time difference, and the formula is expressed as:

[0104] ;

[0105] Where, 、 are the expected start time of the i-th delivery task and the j-th delivery task, respectively, where the i-th delivery task and the j-th delivery task constitute a task pair; The maximum expected start time difference between all delivery task pairs; is the spatial distance between the i-th delivery task and the j-th delivery task (i.e., the shortest path distance from the starting point to the end point of the two delivery tasks); is the maximum spatial distance between all delivery task pairs; 、 The estimated processing time (including running time and loading and unloading time) of the i-th delivery task and the j-th delivery task respectively; is the maximum processing time difference between all delivery task pairs; 、 、 are the weight coefficients of the three dimensions of time difference, spatial distance and processing time.

[0106] The task priority is determined by the ratio of the urgency and conflict of the delivery task, and the formula is expressed as:

[0107] ;

[0108] Where, is the priority value of the i-th delivery task; is the sum of the conflict degrees between the i-th delivery task and all other delivery tasks (i.e., the sum of the pairwise conflict degrees between the i-th delivery task and the remaining n-1 delivery tasks). This formula indicates that the priority of a delivery task is proportional to its urgency and inversely proportional to its total conflict degree. That is, for the same urgency, the delivery task with the least conflict with other delivery tasks will receive a higher priority.

[0109] S203: Execute a multi-population co-evolution algorithm based on the objective function set and the task priority.

[0110] The multi-population co-evolution algorithm includes: constructing K sub-populations; calculating the population diversity index for each sub-population; selecting to perform cross-population recombination or adaptive mutation operations according to the size of the population diversity index; exchanging information between populations every preset generation and updating the global non-dominated solution set.

[0111] Specifically, first construct K subpopulations:

[0112] ;

[0113] in, is the number of subpopulations;

[0114] For each of the subpopulations , when its optimal solution has not been updated in N consecutive generations, calculate the population diversity index:

[0115] ;

[0116] in, is the size of the i-th subpopulation; For subpopulation individuals in is the number of objective functions; is the evaluation value of the j-th objective function on individual x; is the jth objective function in the population The average value in is the diversity index value of the i-th subpopulation;

[0117] like , then perform cross-population recombination:

[0118] ;

[0119] in, are individuals from the current population, are individuals from non-dominated solution sets of other populations; It is a new individual produced after recombination; is the reorganization weight coefficient; is the diversity index value of the i-th subpopulation.

[0120] like , then perform adaptive mutation:

[0121] ;

[0122] in, , g is the current algebra, is the maximum algebra, b is the shape parameter; is the new individual after mutation; x is the individual before mutation; is the upper bound of the decision variable; is the lower bound of the decision variable; is a random number, ∈[0,1]; Adaptive variable step length.

[0123] Information exchange is performed between populations every M generations, and the global non-dominated solution set is updated by calculating the dominance relationship of the solutions.

[0124] S204: Calculate a comprehensive score for each solution based on the global non-dominated solution set and the task priority; select the optimal solution according to the comprehensive score to generate the first AGV delivery task sequence, where the first AGV delivery task sequence includes AGV number, task number, and execution time information.

[0125] S3: Based on the material demand forecast result and the first AGV delivery task sequence, a second AGV delivery task sequence is generated, and the AGV is controlled to perform the material delivery task according to the second AGV delivery task sequence, while task execution monitoring is performed.

[0126] Specifically, monitoring includes AGV location tracking, task completion evaluation, and abnormal situation reporting.

[0127] In summary, the present invention realizes intelligent optimization of material distribution management by combining LSTM network for material demand forecasting, adopting multi-objective optimization algorithm for task sequence generation and real-time monitoring of AGV execution status, significantly improving the operating efficiency of AGV distribution system, reducing energy consumption and achieving dynamic balancing of equipment load.

[0128] Example 2 is an embodiment of the present invention, which provides an AGV-based intelligent material distribution management optimization system, including: a data analysis module, used to acquire and analyze factory operation data to obtain material demand forecast results; a task generation module, which generates a first AGV distribution task sequence based on factory operation data using a multi-objective optimization algorithm; a task optimization execution module, which generates a second AGV distribution task sequence based on the material demand forecast results and the first AGV distribution task sequence, controls the AGV to perform material distribution tasks according to the second AGV distribution task sequence, and monitors the task execution at the same time.

[0129] Example 3, reference Figure 2 , which is an embodiment of the present invention, differs from the previous embodiment in that: if the functions described are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0130] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0131] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, and then editing, interpreting, or processing in another suitable manner as necessary, and then storing it in a computer memory.

[0132] It should be understood that various components of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof may be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0133] Example 4 is an embodiment of the present invention, which provides an AGV-based intelligent material distribution management optimization method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0134] Experimental verification was conducted at a large automotive parts manufacturing plant with 42 production workstations and 12 AGV robots. Production data from October 2023 to March 2024 was selected as the experimental baseline data. First, RFID data collection terminals deployed at each workstation collected material demand data for six months, including 84,632 delivery records for 314 different types of materials. A data preprocessing module was used to clean the raw data, removing outliers and obtaining 82,947 valid data points. Subsequently, the data was divided into three categories based on material type: A, B, and C. Category A comprises high-frequency critical materials (accounting for 30% of total demand), Category B comprises medium-frequency materials (accounting for 45% of total demand), and Category C comprises low-frequency materials (accounting for 25% of total demand).

[0135] Using a modified autocorrelation function to analyze the demand cycle characteristics of various types of materials, Class A materials primarily exhibit dual cycles of 4 and 8 hours, Class B materials exhibit a single cycle of 12 hours, and Class C materials exhibit a long cycle of 24 hours. Based on these cycle characteristics, a three-layer LSTM network (128 input layer nodes, 256 hidden layer nodes, and 64 output layer nodes) was constructed. Data from the first five months was used for training, and the last month's data was used for testing. Simultaneously, the AGV's onboard sensor system collected real-time position, power, and load data at a sampling frequency of 1 Hz to construct an AGV status database.

[0136] In the task optimization phase, four objective functions were designed: a distribution efficiency objective function (optimizing the shortest completion time), an energy consumption objective function (optimizing the minimum total energy consumption), an equipment load balancing objective function (optimizing the minimum standard deviation of each AGV's workload), and an urgency objective function (optimizing the highest priority for urgent tasks). An improved multi-population co-evolutionary algorithm was used, with a population size of 200, 1000 generations, a crossover probability of 0.8, and a mutation probability of 0.1. In the experiment, the traditional method, the single-objective optimization method, and the method of the present invention were used for comparative tests. Each test was repeated 30 times and the average value was taken. The results are shown in Table 1:

[0137] Table 1 Experimental data comparison table:

[0138] Optimization methods Average delivery completion time (min) Energy consumption index (kWh / h) Device load balancing degree (%) Emergency task response time (min) Material delivery accuracy (%) System suitability score Traditional rule-based approach 45.8 12.4 68.5 15.6 92.3 0.72 Single-objective optimization method 38.2 10.8 72.4 12.3 94.8 0.78 Group A of the present invention 28.4 8.6 89.7 6.8 98.2 0.91 Group B of the present invention 29.1 8.9 88.5 7.2 97.8 0.90 Group C of the present invention 27.9 8.4 90.2 6.5 98.5 0.92 Group D of the present invention 28.6 8.7 89.1 6.9 98.1 0.91 ;

[0139] Comparing experimental data reveals the following analytical conclusions: The present invention significantly outperforms traditional methods and single-objective optimization methods across all performance indicators. In terms of average delivery completion time, the present invention reduces delivery time by approximately 39.1% compared to traditional rule-based methods and by approximately 25.7% compared to single-objective optimization methods. This significant improvement is primarily due to the LSTM network's precise prediction of material demand and the synergistic effect of the multi-objective optimization algorithm. Regarding energy consumption, the present invention achieves an average energy consumption of 8.65 kWh / h, a 30.2% reduction compared to traditional methods. This is attributed to the comprehensive consideration of AGV energy consumption during task sequence optimization.

[0140] In terms of equipment load balancing, the proposed method achieved an average of 89.4%, a 30.5 percentage point improvement over traditional methods, demonstrating the superiority of multi-objective optimization in task allocation. Regarding emergency task response time, the proposed method achieved an average response time of only 6.85 minutes, a 56.1% improvement over traditional methods, thanks to a dynamic adjustment mechanism for task priorities. Material delivery accuracy reached a high of 98.2%, a 5.9 percentage point improvement over traditional methods, demonstrating the high precision of the prediction model.

[0141] The system adaptability score is a comprehensive evaluation indicator that includes multiple dimensions such as the system's ability to handle emergencies and its ability to dynamically adjust tasks. The average score of the present invention is 0.91, which is 26.4% higher than that of traditional methods, indicating that this method has stronger environmental adaptability and robustness. Through variance analysis of the four groups of experimental data A, B, C, and D, when the confidence level is 95%, the standard deviation of each indicator is less than 3%, proving that the present invention has good stability and repeatability. These experimental data fully demonstrate the innovation and practical value of the present invention in the field of intelligent material distribution management optimization.

[0142] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An AGV-based intelligent material distribution management optimization method, characterized in that: include: Acquire and analyze factory operation data to obtain material demand forecast results; the factory operation data includes material demand data and AGV status data; Based on the factory operation data, a first AGV delivery task sequence is generated using a multi-objective optimization algorithm; Based on the material demand forecast result and the first AGV delivery task sequence, a second AGV delivery task sequence is generated, and the AGV is controlled to perform the material delivery task according to the second AGV delivery task sequence, while monitoring the task execution; The process of obtaining the material demand forecast result includes the following steps: Classifying the material demand data according to material types to obtain material classification data; extracting material demand cycle characteristics from the material classification data; Based on the material demand cycle characteristics, a long short-term memory network (LSTM) is used for training to obtain a material demand forecasting model; Calculating AGV delivery capability data based on the AGV status data; Generate material demand forecast results based on the material demand forecast model, material cycle characteristics and AGV delivery capacity data; The method of generating a first AGV delivery task sequence based on factory operation data using a multi-objective optimization algorithm includes the following steps: constructing a set of objective functions based on the plant operation data; Determining the degree of conflict between AGV delivery tasks based on the objective function set, and determining the task priority according to the conflict degree; executing a multi-population co-evolution algorithm based on the objective function set and the task priorities; Calculating a comprehensive score for each solution based on a global non-dominated solution set and the task priority; selecting an optimal solution according to the comprehensive score to generate the first AGV delivery task sequence, wherein the first AGV delivery task sequence includes AGV number, task number, and execution time information; The objective function set includes a distribution efficiency objective function, an energy consumption objective function, an equipment load balancing objective function and an urgency objective function; The conflict degree is calculated comprehensively by time difference, spatial distance and processing time difference, and the formula is expressed as follows: ； Where, 、 are the expected start time of the i-th delivery task and the j-th delivery task, respectively, where the i-th delivery task and the j-th delivery task constitute a task pair; The maximum expected start time difference between all delivery task pairs; is the spatial distance between the i-th delivery task and the j-th delivery task; is the maximum spatial distance between all delivery task pairs; 、 are the estimated processing time of the i-th delivery task and the j-th delivery task respectively; is the maximum processing time difference between all delivery task pairs; 、 、 are the weight coefficients of the three dimensions of time difference, spatial distance and processing time; The task priority is determined by the ratio of task urgency to conflict degree: ； Where, is the priority value of the i-th delivery task; is the sum of the conflict degrees between the ith delivery task and all other delivery tasks; u i The urgency of the task.

2. The AGV-based intelligent material distribution management optimization method according to claim 1, characterized in that: The material requirement data includes material type, quantity and urgency; The AGV status data includes current position, power level and load status; The process of extracting the material demand cycle characteristics is specifically to analyze the material classification data through the autocorrelation function, find the main cycle, and construct a cycle characteristic function, and obtain the material cycle characteristics based on the cycle characteristic function.

3. The AGV-based intelligent material distribution management optimization method according to claim 2, characterized in that: The multi-population co-evolution algorithm includes: Construct K subpopulations: ； in, is the number of subpopulations; For each subpopulation, when its optimal solution has not been updated for N consecutive generations, calculate: ； in, is the size of the i-th subpopulation; For subpopulation individuals in is the number of objective functions; is the evaluation value of the kth objective function on individual x; is the kth objective function in the population The average value in is the diversity index value of the i-th subpopulation; like , then perform cross-population recombination: ； in, are individuals from the current subpopulation, are individuals in the non-dominated solution set from other subpopulations; It is a new individual produced after recombination; is the reorganization weight coefficient; is the threshold of the diversity indicator; like , then perform adaptive mutation: ； in, , g is the current algebra, is the maximum algebra, b is the shape parameter; is the new individual after mutation; x is the individual before mutation; is the upper bound of the decision variable; is the lower bound of the decision variable; is a random number, ∈[0,1]; To adaptively change the asynchronous length; Information exchange is performed between populations every M generations, and the global non-dominated solution set is updated by calculating the dominance relationship of the solutions.

4. The AGV-based intelligent material distribution management optimization method according to claim 3, characterized in that: The monitoring includes AGV location tracking, task completion assessment and abnormal situation reporting.

5. A system using the AGV-based intelligent material distribution management optimization method according to any one of claims 1 to 4, characterized in that: include: Data analysis module, used to obtain and analyze factory operation data to obtain material demand forecast results; a task generation module, which generates a first AGV delivery task sequence based on the factory operation data using a multi-objective optimization algorithm; A task optimization execution module generates a second AGV delivery task sequence based on the material demand forecast result and the first AGV delivery task sequence, controls the AGV to execute the material delivery task according to the second AGV delivery task sequence, and monitors the task execution at the same time.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the AGV-based intelligent material distribution management optimization method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the AGV-based intelligent material distribution management optimization method according to any one of claims 1 to 4 are implemented.

Images