Automatic wharf AGV intelligent scheduling method and system

Abstract

The invention relates to the technical field of intelligent scheduling, and discloses an automatic wharf AGV intelligent scheduling method and system. The method comprises the steps of collecting container task data through a wharf operation management system, performing multi-dimensional evaluation processing to obtain a task priority evaluation matrix, constructing an AGV state vector according to the matrix, performing load capacity prediction based on historical data, and performing intelligent matching on the task matrix and the AGV state vector through a Port-MTCSA algorithm to obtain a whole-process task chain. And performing time sequence analysis based on the task chain to predict wharf load distribution, and executing AGV resource dynamic re-planning when the load exceeds a threshold value. The technical problem that an existing automatic wharf AGV scheduling method lacks dynamic priority evaluation, load prediction, task chain optimization and prospective resource allocation in a multi-task concurrent scene is solved.

Description

Technical Field

[0001] This application relates to the field of intelligent scheduling technology, and particularly to an intelligent scheduling method and system for AGVs in automated terminals. Background Art

[0002] The existing AGV scheduling methods in automated terminals mainly adopt static path planning combined with simple conflict avoidance strategies, and manage the container transportation operations in the terminal by pre-computing the running paths and task assignment schemes of AGVs. Traditional scheduling methods usually perform task assignment based on fixed priority rules and greedy algorithms, and use simple strategies such as first-come-first-served or shortest-distance-first to determine the operation order of AGVs. At the same time, time window constraints and capacity limits are adopted to avoid resource conflicts. These methods can maintain basic scheduling functions when the terminal operations are relatively stable and the task volume is small, providing technical support for the initial development of automated terminals.

[0003] However, the existing technologies have significant deficiencies in dealing with multi-task concurrent scenarios, mainly reflected in the lack of an intelligent evaluation and real-time adjustment mechanism for dynamic task priorities, and the inability to re-order tasks according to the actual urgency and business value of terminal operations. Most of the existing task assignment algorithms adopt greedy or simple heuristic strategies, and it is difficult to achieve global optimal assignment under multi-objective constraints. Especially when considering multiple constraints such as time windows, AGV energy consumption, and equipment status simultaneously, static scheduling schemes often lead to uneven resource utilization and low operation efficiency. In addition, traditional methods lack predictive scheduling capabilities based on historical operation data, and are unable to identify peak-hour task congestion and resource bottlenecks in advance, resulting in passivity and lag in scheduling decisions, and it is difficult to maintain optimal scheduling performance when the terminal operation load changes dynamically.

[0004] Based on an in-depth analysis of the above technical defects, the fundamental problems of the existing technologies in intelligent scheduling decisions can be further identified: First, there is a lack of a dynamic evaluation mechanism for task priorities that integrates multi-dimensional factors, and it is unable to make intelligent trade-offs among multiple dimensions such as the urgency of ship departure, container weight level, and customer priority; second, there is a lack of a load capacity prediction method based on the real-time status and historical operation patterns of AGVs, and it is unable to accurately estimate the changes in the operation capabilities of AGVs within future time windows; third, there is a lack of a specialized task chain optimization algorithm for the three-layer operation structure of "quayside crane - AGV - yard" in the terminal, and it is unable to achieve coordinated optimization of the entire process of loading / unloading - transportation - stacking; fourth, there is a lack of a forward-looking resource allocation mechanism based on load prediction, and it is unable to actively perform dynamic re-allocation of AGV resources before the arrival of the load peak. Summary of the Invention

[0005] This application provides an intelligent scheduling method and system for AGVs in automated terminals, aiming to solve the technical problems that existing AGV scheduling methods in automated terminals lack dynamic priority evaluation, load prediction, task chain optimization, and forward-looking resource allocation in multi-task concurrent scenarios.

[0006] In the first aspect, this application provides an intelligent scheduling method for AGVs in automated terminals. The intelligent scheduling method for AGVs in automated terminals includes: collecting container task data through the terminal operation management system, and performing multi-dimensional evaluation and processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix; constructing an AGV state vector based on the task priority evaluation matrix, and performing load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with port historical operation data to obtain an AGV load prediction result; matching the AGV load prediction result with the task priority evaluation matrix, and performing intelligent matching processing on the container task priority matrix and the AGV dynamic load status vector through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a full-process task chain for loading, transporting, and storing; collecting terminal historical operation data according to the full-process task chain for loading, transporting, and storing, and performing load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the terminal operation load distribution; comparing the terminal operation load distribution with a preset load threshold, and when the predicted load exceeds the threshold, performing dynamic re-planning processing on the AGV resource allocation to obtain an optimized scheduling instruction and execute the intelligent scheduling of AGVs in the terminal.

[0007] In the second aspect, this application provides an intelligent scheduling system for AGVs in automated terminals. The intelligent scheduling system for AGVs in automated terminals includes: An evaluation module, configured to collect container task data through the terminal operation management system, and perform multi-dimensional evaluation and processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix; A prediction module, configured to construct an AGV state vector based on the task priority evaluation matrix, and perform load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with port historical operation data to obtain an AGV load prediction result; A matching module, configured to match the AGV load prediction result with the task priority evaluation matrix, and perform intelligent matching processing on the container task priority matrix and the AGV dynamic load status vector through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a full-process task chain for loading, transporting, and storing; A collection module, configured to collect historical terminal operation data according to the entire loading-unloading-transportation-storage process task chain, and perform load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the terminal operation load distribution; A comparison module, configured to compare the terminal operation load distribution with a preset load threshold, and when the predicted load exceeds the threshold, perform dynamic re-planning processing on the AGV resource allocation to obtain an optimized scheduling instruction and execute the intelligent scheduling of the terminal AGV.

[0008] In a third aspect, an intelligent scheduling device for an automated terminal AGV is provided, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory to enable the intelligent scheduling device for the automated terminal AGV to execute the above-mentioned intelligent scheduling method for the automated terminal AGV.

[0009] In a fourth aspect, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and when it runs on a computer, it enables the computer to execute the above-mentioned intelligent scheduling method for the automated terminal AGV.

[0010] In the technical solution provided by this application, by establishing a multi-dimensional evaluation and processing mechanism based on the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority, the problem of unreasonable resource allocation caused by static task allocation in the prior art is solved. The task priority evaluation matrix can dynamically adjust the task execution order according to the actual situation of terminal operations, avoiding the delay of urgent tasks caused by the traditional first-come-first-served strategy. The construction of the AGV state vector and the load capacity prediction processing based on the historical operation data of the port overcome the defect of ignoring the dynamic state changes of AGVs in the prior art. Through the comprehensive analysis of the current position coordinates, load status, remaining power, and estimated completion time of the AGV, an accurate prediction of the future operation ability of the AGV is realized, significantly improving the accuracy and executability of task allocation. The terminal multi-task chain scheduling algorithm Port-MTCSA realizes the systematic optimization of the entire loading-unloading-transportation-storage process task chain through the intelligent matching of the container task priority matrix and the AGV dynamic load status vector. Compared with the traditional segmented scheduling method, the entire process task chain can comprehensively consider the coordination of the entire operation link, reduce the waiting time and resource idleness between operation links, and improve the overall operation efficiency of the terminal.

[0011] Based on the time - series analysis of load prediction, through the comprehensive analysis of ship arrival plans, weather factors, and container type distributions, a forward - looking terminal operation load prediction mechanism is established, which solves the problem of passive response to load changes in the existing technology, enabling the AGV scheduling to actively adjust resource allocation before the arrival of the load peak. The dynamic re - planning processing mechanism compares the terminal operation load distribution with a preset load threshold and automatically triggers the re - allocation of AGV resources when the predicted load exceeds the threshold, overcoming the deficiency of the traditional scheduling method lacking adaptive adjustment ability and realizing the dynamic optimization configuration of terminal AGV resources. Especially in the specific application field of AGV intelligent scheduling in automated terminals, the Port - MTCSA algorithm is specifically designed for the particularity of the three - layer operation structure of "quayside crane - AGV - yard" in the terminal. The three - layer network modeling and dynamic programming recurrence characteristics of the algorithm enable it to effectively handle the complex constraints of terminal operations. Compared with the general vehicle routing problem - solving algorithms, the computational efficiency and optimization effect of Port - MTCSA in the terminal environment are significantly improved. The introduction of time - series analysis and load prediction algorithms transforms the scheduling decision from passive response to active prevention, and the overall solution achieves a technological leap from static scheduling to intelligent dynamic scheduling, providing key technical support for the efficient operation of modern automated terminals. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Figure 1 FIG. is a schematic diagram of an embodiment of the AGV intelligent scheduling method for an automated terminal in an embodiment of the present application; Figure 2 FIG. is a schematic diagram of an embodiment of the AGV intelligent scheduling system for an automated terminal in an embodiment of the present application; Figure 3 FIG. is a schematic block diagram of the structure of the AGV intelligent scheduling device for an automated terminal in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] The embodiments of the present application provide an intelligent scheduling method and system for AGV in an automated terminal. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims and the above-mentioned drawings of the present application are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the term "comprising" or "having" and any variation thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0014] For ease of understanding, the specific process of the embodiments of the present application will be described below. Please refer to Figure 1 , an embodiment of the intelligent scheduling method for AGV in the automated terminal in the embodiments of the present application includes: Step S101: Collect container task data through the terminal operation management system, and perform multi-dimensional evaluation processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix; Step S102: Construct an AGV state vector according to the task priority evaluation matrix, and perform load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with the port historical operation data to obtain an AGV load prediction result; Step S103: Match the AGV load prediction result with the task priority evaluation matrix, and perform intelligent matching processing on the container task priority matrix and the AGV dynamic load status vector through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a full-process task chain for loading, transporting, and stacking; Step S104: Collect the terminal historical operation data according to the full-process task chain for loading, transporting, and stacking, and perform load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the terminal operation load distribution; Step S105: Compare the terminal operation load distribution with a preset load threshold. When the predicted load exceeds the threshold, perform dynamic re-planning processing on the AGV resource allocation to obtain an optimized scheduling instruction and execute the intelligent scheduling of the terminal AGV.

[0015] It can be understood that the execution subject of the present application can be an intelligent scheduling system for AGV in an automated terminal, or a terminal or a server. Specifically, it is not limited here. The embodiments of the present application will be described by taking the server as the execution subject as an example.

[0016] Specifically, through the data interface with the port informatization platform, basic information such as container numbers, ship voyages, quay crane operation plans, yard locations, and estimated operation times is collected. These raw data are subjected to structured parsing and processing, and data in different formats are uniformly converted into a standardized task data set. Then, the urgency coefficient is calculated based on the difference between the ship's departure time and the current time. The specific calculation method is to convert the time difference into an urgency score, with a higher score for a shorter time. At the same time, the weight parameter in the container specification information is extracted, and 20-foot and 40-foot containers are classified and coded to form a weight grade coefficient matrix. Then, through a multi-dimensional weight distribution algorithm, the urgency score vector, weight grade coefficient matrix are weighted and fused with the quay crane operation sequence, yard distance coefficient, and customer priority to calculate the comprehensive priority value of each task, forming a task priority evaluation matrix.

[0017] According to the task distribution in the task priority evaluation matrix, the real-time operation parameters of each AGV are collected, including the current position coordinates obtained by GPS positioning, the load status detected by the load sensor, the remaining battery power feedback by the battery management system, and the estimated completion time calculated based on the current task. These four parameters are vectorized and encoded to form an AGV status vector. Then, an associated query is performed with the terminal historical operation database, and the AGV load change data within the past 30 days are filtered out by setting a time window. Feature extraction is performed on these historical data to analyze the load distribution law of AGV in different yard areas. Through load frequency calculation, the load probabilities of 20-foot and 40-foot containers in each time period are statistically calculated to generate a load probability distribution table. The current load status in the AGV status vector is matched and calculated with the load probability distribution table to analyze the difference between the remaining load capacity of the AGV and the weight requirement of the task to be executed, obtaining a load matching degree coefficient. Finally, an associated operation is performed in combination with the estimated completion time of the AGV, and the change trend of the load capacity of the AGV within the future time window is predicted through load time series analysis.

[0018] Fuse the AGV load prediction results with the task priority evaluation matrix, calculate the matching degree between the container task weight requirement and the AGV load capacity, and generate an initial task-AGV matching table. Based on this matching table, construct a bipartite graph structure. The Port-MTCSA algorithm for dock multi-task chain scheduling decomposes the dock operation into three levels: quay crane loading and unloading nodes, AGV transportation nodes, and yard storage nodes, and conducts network modeling on the connection relationships between these three levels to form a dock operation network topology graph. Extract the node connection relationships from the network topology graph, chain and arrange the complete operation process of the container from the quay crane to the yard in the order of loading-unloading - transportation - storage to generate multiple candidate task chains. The Port-MTCSA algorithm calculates the cumulative cost of each task chain from the starting state to the ending state through dynamic programming recursion, including time cost, load cost, and distance cost, and obtains the task chain optimization score by minimizing the path cost. Sort and screen the candidate task chains according to the optimization score, adopt the greedy selection strategy to preferentially allocate the high-score task chains, and finally form the optimal matching results of the container task priority matrix and the AGV dynamic load status vector.

[0019] According to the time node information in the full-process task chain of loading-unloading - transportation - storage, extract data such as ship arrival time, operation completion time, and AGV operation trajectory within the past 90 days from the dock historical operation database. Arrange these historical data in a sequence according to the time dimension, identify the periodic laws and seasonal change patterns of the ship arrival plan through time series analysis methods, and establish a ship arrival time series model. At the same time, collect weather forecast data and historical weather records, analyze the correlation between wind speed, rainfall, visibility and dock operation efficiency, and calculate the weather impact coefficient matrix through correlation analysis. Extract container transportation records from the historical operation dataset, and count the quantity distribution of 20-foot and 40-foot containers in different time periods to form the container type distribution law. Integrate the ship arrival time series model, the weather impact coefficient matrix, and the container type distribution law for comprehensive modeling, and calculate the task density distribution of each operation area within the future time window through the load prediction algorithm.

[0020] Compare the load values of each operation area in the terminal operation load distribution with the preset load threshold one by one. When the load value of a certain area exceeds the threshold, mark and identify it to form a list of overloaded operation areas. According to the location information and the degree of load exceeding of the overloaded areas, analyze the allocation status of the current AGV resources in each area, calculate the shortage and surplus of AGV numbers in each area, and generate an AGV resource reallocation demand matrix. The dynamic re-planning algorithm calculates the optimal transfer path and transfer quantity of AGVs from low-load areas to high-load areas according to the demand matrix, considering the AGV movement time cost and fuel consumption, and optimizes the resource adjustment plan. Convert the AGV resource adjustment plan into a specific scheduling instruction format including task assignment, path planning and operation timing, and send the optimized scheduling instructions to each AGV through the terminal 5G communication network to execute intelligent scheduling.

[0021] For example, there are 3 cargo ships berthing at a terminal at the same time. Ship A will depart in 6 hours, Ship B in 12 hours, and Ship C in 24 hours. According to the calculation of time urgency, the task of Ship A gets the highest priority score. There are currently 15 AGVs in operation, 5 of which are in the 1st yard with an empty load status, 8 are in the quay crane area loading 40-foot containers, and 2 are in the 2nd yard in the charging state. Through load prediction analysis, the empty AGVs are suitable for taking on new transportation tasks, while the AGVs that are loading need to complete the current tasks first. The Port-MTCSA algorithm preferentially assigns the urgent task of Ship A to the empty AGVs and at the same time plans the shortest path from the quay crane to the designated yard. Through time series analysis, it is found that the afternoon period is usually the peak operation period. Combining with the weather forecast showing no adverse weather impact, it is predicted that the load in the quay crane area will reach the upper limit of the threshold within the next 2 hours. The dynamic re-planning algorithm decides to deploy 2 fully charged AGVs to the quay crane area and pre-deploy 1 empty AGV to the 2nd yard to prepare for receiving tasks. Through this predictive scheduling, the emergence of operation bottlenecks is avoided.

[0022] In a specific embodiment, the process of executing step S101 may specifically include the following steps: Collect the container number, ship information, quay crane operation plan, yard location and estimated operation time through the data interface of the terminal operation management system, and perform structured parsing and processing on the container task data to obtain a standardized task data set; Calculate the difference between the ship departure time in the standardized task data set and the current time, and perform quantization assignment processing on the urgency coefficient based on the urgency degree of the ship departure time to obtain an urgency score vector; Extract the weight parameter according to the container specification information in the standardized task data set, and perform classification coding processing on the load levels of 20-foot containers and 40-foot containers to obtain a weight level coefficient matrix; The urgency score vector is weighted and fused with the weight level coefficient matrix, and through a multi-dimensional weight distribution algorithm, the quay crane operation sequence, yard distance coefficient, and customer priority are comprehensively calculated and processed to obtain a task priority evaluation matrix.

[0023] Specifically, the terminal operation management system obtains raw data from the port informatization platform through a standardized data interface, including multi-source heterogeneous data such as container numbers, ship voyage information, quay crane operation plans, yard location coordinates, and estimated operation times. The structured parsing process first performs format conversion on different formats of data, unifying the conversion of XML-format ship information, JSON-format quay crane plans, and database-format yard data into a standard data structure. During the parsing process, key fields are extracted. The container number is converted into a unique identifier, and the ship information is parsed to obtain attributes such as ship name, voyage, and estimated departure time. The quay crane operation plan is parsed to obtain the operation serial number, assigned quay crane number, and estimated loading and unloading time. The yard location is parsed to obtain the specific yard area number and storage location coordinates. The standardized task data set organizes all task information according to unified fields such as task ID, ship ID, container specification, starting position, target position, and time window, forming a structured data table, where each row represents a specific container transportation task and each column represents an attribute dimension of the task.

[0024] The quantization assignment process of the urgency coefficient is based on two core steps: time difference calculation and urgency degree mapping. First, the ship departure time corresponding to each task is extracted from the standardized task data set, and the difference is calculated with the current time to obtain the remaining time. The remaining time is quantified in hours, and then the urgency coefficient is assigned through a piecewise function. The urgency coefficient of tasks with remaining time between 0 and 6 hours is set to 0.9 to 1.0, between 6 and 12 hours is set to 0.7 to 0.9, between 12 and 24 hours is set to 0.5 to 0.7, and over 24 hours is set to 0.3 to 0.5. The specific quantization assignment uses the linear interpolation method. Within each time period, the exact urgency coefficient is calculated according to the specific value of the remaining time. The linear interpolation calculation determines the upper and lower bounds of the time period and the corresponding coefficient range, and calculates the accurate coefficient value according to the equal proportion relationship. The urgency score vector arranges the urgency coefficients of all tasks in the order of tasks, forming a one-dimensional array structure. Each element in the vector corresponds to the urgency coefficient value of a task, and the length of the vector is equal to the total number of tasks.

[0025] The extraction of weight parameters starts from the container specification information field in the standardized task dataset. This field contains the size specifications and load information of the containers. Through classification and coding, the containers are divided into two main categories: 20-foot and 40-foot. Each category is further subdivided according to the load capacity. The 20-foot containers are divided into three grades according to the load range: light load, medium load, and heavy load. The corresponding load ranges are 0 to 10 tons, 10 to 15 tons, and 15 to 24 tons respectively, and the coding values are set as 1, 2, and 3. The 40-foot containers are also divided into three load grades, with load ranges of 0 to 15 tons, 15 to 25 tons, and 25 to 30 tons respectively, and the coding values are set as 4, 5, and 6. The weight grade coefficient matrix adopts a two-dimensional matrix structure. The rows represent the task numbers, and the columns represent the weight grade attributes. The values in the matrix are the corresponding weight grade codes. During the matrix construction process, the weight grade code of each task is determined according to the container specification information of the task, and the coding value is filled into the corresponding position in the matrix. The number of rows in the matrix is equal to the total number of tasks, and the number of columns is fixed at 1 column to store the weight grade code.

[0026] The multi-dimensional weight assignment algorithm performs weighted fusion calculations on the urgency score vector, weight grade coefficient matrix, quay crane operation sequence, yard distance coefficient, and customer priority. The quay crane operation sequence is numbered according to the operation plan order of the quay crane. The earlier the task sequence number, the higher the weight. The sequence weight is calculated using the reciprocal relationship. The sequence weight of the first task is 1.0, the second is 0.5, the third is 0.33, and so on. The yard distance coefficient is obtained by calculating the straight-line distance from the quay crane position to the target yard position. The closer the distance, the higher the weight. The distance weight is calculated using an inverse proportion relationship, specifically, 1000 divided by the distance in meters and then divided by 1000 to obtain the standardized weight value. The customer priority is comprehensively evaluated based on factors such as the customer's VIP level, cooperation years, and business volume, and is divided into three grades: high, medium, and low, with corresponding weight coefficients of 1.0, 0.7, and 0.4 respectively. The weight fusion calculation adopts the linear weighted summation method and is calculated through the following formula: M = G×0.3 + H×0.2 + J×0.2 + K×0.15 + L×0.15 + O×0.15, where M is the comprehensive priority score, G is the urgency coefficient, H is the weight grade coefficient, J is the operation sequence coefficient, K is the distance coefficient, and O is the customer priority coefficient. According to the comprehensive priority score obtained for each task, the task priority evaluation matrix organizes the comprehensive priority scores of all tasks in matrix form. The rows represent the tasks, and the columns represent the priority scores. The values in the matrix reflect the importance of each task in the scheduling.

[0027] In a specific embodiment, the process of executing step S102 may specifically include the following steps: Collect the real-time operation parameters of the AGV according to the task distribution in the task priority evaluation matrix, perform vectorized encoding processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV to obtain the AGV status vector; Associate and query the AGV status vector with the terminal historical operation database, and perform feature extraction processing on the AGV load change data within the past 30 days through time window screening to obtain the historical load feature sample set; Based on the historical load feature sample set, statistically analyze the load distribution of the AGV in different yard areas, and perform quantitative processing on the load probabilities of 20-foot and 40-foot containers through load frequency calculation to obtain the load probability distribution table; Perform matching calculation according to the current load status in the AGV status vector and the load probability distribution table, and perform difference analysis processing on the remaining load capacity of the AGV and the weight requirement of the task to be executed to obtain the load matching degree coefficient; Perform an association operation between the load matching degree coefficient and the estimated completion time of the AGV, and perform prediction processing on the change trend of the load capacity of the AGV within the future time window through load time series analysis to obtain the AGV load prediction result.

[0028] Specifically, collect the real-time operation parameters of the AGV according to the task distribution in the task priority evaluation matrix, and obtain the real-time status information of each AGV through the communication interface with the AGV vehicle control system. Among them, the current position coordinates are obtained through the fusion positioning of the GPS positioning module and the terminal internal positioning system, the load status is to detect the weight of the currently loaded container in real time through the vehicle-mounted weighing sensor, the remaining power is to read the current battery power percentage through the battery management system, and the estimated completion time is calculated according to the remaining path of the currently executing task and the standard driving speed of the AGV. The vectorized encoding processing converts these four real-time parameters into the form of a numerical vector. The position coordinates are represented by the X-axis and Y-axis values of the terminal coordinate system, the load status is represented by the ratio of the current load to the maximum load capacity of the AGV, the remaining power is represented by a decimal between 0 and 1, and the estimated completion time is represented by an integer in minutes. The AGV status vector arranges these four numerical values in a fixed order to form a four-dimensional vector, and each component of the vector corresponds to a status parameter.

[0029] Perform an associated query on the AGV status vector and the terminal historical operation database, and extract the AGV load change records within the past 30 days from the historical database through a time window filtering mechanism. The time window filtering conducts a range query based on the timestamp field of the data records, and the filtering condition is that the record time is greater than the current time minus 30 days and less than the current time. Feature extraction processing extracts key features such as AGV number, timestamp, location information, load status, and task type from the filtered historical records. The feature extraction uses the attribute selection method in data mining to determine the most predictive feature dimensions based on the correlation analysis of load prediction. The historical load feature sample set organizes the extracted feature data in the form of a two-dimensional matrix of sample-feature. Each row represents the AGV status sample at a historical moment, and each column represents a feature dimension. The scale of the sample set depends on the total number of data records within 30 days and the number of feature dimensions.

[0030] Based on the historical load feature sample set, conduct a statistical analysis on the load distribution of AGV in different yard areas. The statistical analysis uses the grouped statistical method, grouping the historical samples by yard area. Each yard area contains the historical load records of all AGVs within that area. The load frequency calculation statistically analyzes the occurrence times and load distribution of 20-foot containers and 40-foot containers within each yard area. During the frequency calculation process, the 20-foot containers are classified and statistically analyzed according to three load levels: light load, medium load, and heavy load. The 40-foot containers are also classified according to three load levels. The load probability quantization process divides the occurrence times of each level by the total number of load records in that area to obtain the probability value. The load probability distribution table adopts a three-dimensional table structure. The first dimension is the yard area, the second dimension is the container type, and the third dimension is the load level. The values in the table are the corresponding probability values.

[0031] Perform a matching calculation based on the current load status in the AGV status vector and the load probability distribution table. The matching calculation first determines the yard area where the AGV is located according to its current position, and then queries the load probability distribution of that area from the load probability distribution table. The difference analysis processing calculates the difference between the remaining load capacity of the AGV and the weight requirement of the task to be executed. The remaining load capacity is equal to the maximum load capacity of the AGV minus the current load. The weight requirement of the task to be executed extracts the container weight information of the corresponding task from the task priority evaluation matrix. The load matching degree coefficient is obtained by normalizing the ratio of the remaining load capacity to the task weight requirement. When the remaining load capacity is greater than the task weight requirement, the matching degree coefficient is close to 1. When the remaining load capacity is insufficient, the matching degree coefficient is close to 0. The matching degree calculation uses the sigmoid function for non-linear mapping to convert the difference into a matching degree coefficient between 0 and 1.

[0032] Perform an association operation on the load matching degree coefficient and the expected completion time of the AGV. The association operation establishes a composite evaluation model of the load matching degree and time factors. The load time series analysis uses the time series prediction method to analyze the changing law of the AGV load capacity over time. The time series analysis is based on the time series data in the historical load feature sample set, and identifies the trend and periodic pattern of the load change through the moving average method. The prediction process calculates the load state of the AGV after completing the current task within the future time window. The prediction calculation considers the load change during the task execution process, including the increase in load when loading the container and the decrease in load when unloading the container. The AGV load prediction result includes the expected load state and load capacity availability at each future time point. The prediction result is stored in the form of a two-dimensional array of time-load state, providing load capacity prediction information for subsequent task allocation decisions.

[0033] In a specific embodiment, the process of executing step S103 may specifically include the following steps: Perform data fusion on the AGV load prediction result and the task priority evaluation matrix, perform a matching degree calculation process on the container task weight requirement and the AGV load capacity, and obtain an initial task-AGV matching table; Based on the initial task-AGV matching table, construct a bipartite graph structure, and perform three-layer network modeling processing on the quay crane loading and unloading nodes, AGV transportation nodes, and yard storage nodes through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain the terminal operation network topology diagram; Extract the task chain path according to the node connection relationship in the terminal operation network topology diagram, and perform chain arrangement processing on the loading-unloading-transportation-storage operation sequence of the container from the quay crane to the yard to obtain a candidate task chain set; Perform cumulative cost calculation on the time constraint and load constraint in the candidate task chain set, and minimize the path cost of the task chain from the start state to the end state through dynamic programming recursion to obtain the task chain optimization score; Sort and filter the candidate task chains based on the task chain optimization score, and perform priority allocation processing on the high-score task chains through the greedy selection strategy to obtain the matching result of the container task priority matrix and the AGV dynamic load state vector; Perform verification processing on the matching result of the container task priority matrix and the AGV dynamic load state vector, and perform conflict detection processing on the load conflict and time conflict in the task chain to obtain the loading-unloading-transportation-storage full-process task chain.

[0034] Specifically, the AGV load prediction results are fused with the task priority evaluation matrix. In the data fusion process, the AGV load capacity data in the load prediction results is first associated with the task weight requirement data in the task priority evaluation matrix. The fusion algorithm uses matrix operations to perform the Cartesian product operation on the AGV load capacity matrix and the task weight requirement matrix to generate all possible combinations of AGVs and tasks. The matching degree calculation process evaluates the load adaptability of each AGV-task combination. The calculation method is to divide the remaining load capacity of the AGV by the task weight requirement to obtain the load matching ratio. When the ratio is greater than or equal to 1, it indicates that the AGV load capacity is sufficient; when the ratio is less than 1, it indicates that the load capacity is insufficient. The matching degree value is converted into a value between 0 and 1 through logarithmic transformation and normalization of the ratio. The task-AGV initial matching table adopts a two-dimensional table structure, with rows representing task numbers and columns representing AGV numbers. The values in the table are the corresponding matching degree scores. During the table construction process, all calculated matching degree scores are filled into the corresponding positions according to the correspondence between tasks and AGVs.

[0035] Based on the task-AGV initial matching table, a bipartite graph structure is constructed. A bipartite graph is a special graph structure that divides nodes into two non-overlapping sets. One set contains all task nodes, and the other set contains all AGV nodes. Edges are only connected between nodes in the two sets. The quay multi-task chain scheduling algorithm Port-MTCSA performs three-layer network modeling on quay crane loading and unloading nodes, AGV transportation nodes, and yard storage nodes. The three-layer network structure decomposes the quay operation process into three levels. The quay crane loading and unloading node layer contains the loading and unloading operation positions of all quay cranes. The AGV transportation node layer contains all the transportation path nodes of AGVs. The yard storage node layer contains all the storage positions in the yard. The network modeling process constructs the network topology by defining node attributes and connection relationships. Node attributes include parameters such as position coordinates, capacity limits, and operation times. The connection relationships are determined by calculating the distances and connectivity between nodes. The quay operation network topology graph graphically represents all the nodes and connection relationships of the three-layer network. Each node in the graph represents an operation position, and each edge represents a feasible transportation path.

[0036] Extract the task chain path according to the node connection relationship in the terminal operation network topology diagram. The path extraction uses a graph traversal algorithm to search all possible paths from the quay crane node to the yard node. During the path search process, consider the connectivity and capacity constraints of the nodes and exclude unreachable path combinations. The chain arrangement process arranges the operation process of the container from the quay crane to the yard according to the time sequence relationship of loading / unloading - transportation - storage. The loading / unloading stage includes the operation of unloading the container from the ship to the quay crane. The transportation stage includes the operation of the AGV transporting the container from the quay crane to the yard. The storage stage includes the operation of storing the container at the designated yard location. The candidate task chain set contains all task chain paths that meet the operation process requirements. Each task chain contains a sequence of operation nodes and the corresponding AGV allocation scheme. The size of the set depends on the number of feasible paths and the availability of AGVs.

[0037] Calculate the cumulative cost for the time constraint and load constraint in the candidate task chain set. The time constraint includes the operation time of each operation node and the moving time of the AGV between nodes. The load constraint includes the load status of the AGV at each node and the load capacity limit. The cumulative cost calculation uses the path cost accumulation method. Starting from the starting node of the task chain, calculate the operation cost of each node and the moving cost between nodes one by one, and accumulate all the costs to obtain the total cost of the task chain. Dynamic programming recursion is an optimization algorithm that solves complex problems by decomposing them into sub - problems. During the recursion process, maintain a state transition table to record the optimal solution of each sub - problem. The state transition equation defines the cost calculation rule from one state to another. The minimization process selects the path with the minimum cost as the optimal solution by comparing the cumulative costs of all candidate paths. The task chain optimization score is represented in the form of the reciprocal or negative number of the cost. The smaller the cost, the higher the score.

[0038] Sort and filter the candidate task chains based on the optimized score of the task chain. The sorting algorithm uses efficient sorting methods such as quicksort or mergesort to arrange the task chains in descending order of score. The greedy selection strategy is a heuristic optimization method that selects the task chain with the highest score in the current state for allocation each time. The core idea of the greedy strategy is that local optimal selection can lead to a global optimal solution. The priority allocation process allocates high-score task chains in sequence according to the sorting result. During the allocation process, the availability of AGVs and the time window constraints of tasks are checked to ensure the feasibility of the allocation plan. The matching result of the container task priority matrix and the AGV dynamic load status vector records the final allocation relationship between each task and the AGV and the corresponding score information. Verify the matching result of the container task priority matrix and the AGV dynamic load status vector. The verification process checks the rationality and executability of the allocation plan. The load conflict detection checks whether the load capacity of the AGV meets the weight requirements of the allocated tasks, and the time conflict detection checks whether there are overlaps or violations of the time window constraints in the task schedule. The conflict detection and handling adopt the solution method of the constraint satisfaction problem, and identify potential conflicts by checking the satisfaction of all constraint conditions. When a conflict is detected, a reallocation mechanism is triggered to adjust the task allocation plan. The loading-unloading-transportation-storage full-process task chain is the final scheduling plan verified and optimized, which includes the complete operation process of each container task and the corresponding AGV allocation plan.

[0039] In a specific embodiment, the process of performing three-layer network modeling processing on the quay crane loading and unloading node, the AGV transportation node, and the yard storage node through the terminal multi-task chain scheduling algorithm Port-MTCSA may specifically include the following steps: Extract node information based on the task allocation relationship in the task-AGV initial matching table, perform spatial coordinate mapping processing on the quay crane loading and unloading positions, the AGV transportation paths, and the yard storage positions to obtain a three-layer node coordinate set; Divide the three-layer node coordinate set according to the terminal operation process, and perform topological construction processing on the node connection relationships of the quay crane layer, the transportation layer, and the yard layer through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a hierarchical network structure; Construct a connection weight matrix according to the node attribute information in the hierarchical network structure, and perform quantization calculation processing on the distance weight from the quay crane loading and unloading node to the AGV transportation node and the distance weight from the AGV transportation node to the yard storage node to obtain a node connection weight table; Perform correlation analysis on the node connection weight table and the container transportation capacity constraint, and perform differential assignment processing on the transportation path weights of 20-foot and 40-foot containers through the Port-MTCSA algorithm to obtain a capacity constraint weight matrix; Finally, a three - layer network structure is constructed based on the capacity - constrained weight matrix. The connectivity of quay crane handling nodes, AGV transportation nodes, and yard storage nodes is verified through graph - theory algorithms to obtain the topological graph of the terminal operation network.

[0040] Specifically, node information is extracted based on the task - AGV initial matching table's task assignment relationship. In the process of extracting node information, all involved quay crane positions, key points of AGV transportation paths, and yard storage positions are identified from the matching table. The extraction algorithm traverses each task - AGV combination in the matching table, obtains the starting quay crane position and the target yard position from the task information, and obtains the current position and reachable path nodes from the AGV information. The spatial - coordinate mapping process converts the actual positions in the terminal physical layout into a digital coordinate system. The quay crane handling positions are assigned coordinates according to the physical distribution of the terminal quay line, and each quay crane position corresponds to a two - dimensional coordinate point. The AGV transportation path includes all feasible path nodes connecting the quay crane and the yard, and the path - node coordinates are determined according to the actual layout of the terminal road network. The yard storage positions are coordinate - mapped according to the division of the yard area and the container storage positions. The three - layer node - coordinate sets divide all the extracted nodes into three sets according to the functional type. The quay - crane layer set contains the coordinates of all quay crane handling positions, the transportation - layer set contains the coordinates of all AGV path nodes, and the yard - layer set contains the coordinates of all yard storage positions. Each set is sorted in an orderly manner according to the node number inside, and the sets are associated through a hierarchical relationship.

[0041] The three - layer node - coordinate sets are hierarchically divided according to the terminal operation process. The hierarchical division follows the natural process order of container operations. The quay - crane layer, as the first layer, is responsible for container handling operations. The transportation layer, as the middle layer, is responsible for container transportation operations. The yard layer, as the final layer, is responsible for container storage operations. The terminal multi - task chain scheduling algorithm Port - MTCSA constructs the topology of the node connection relationships between the three layers. In the topology - construction process, the mutual relationships of the nodes within the layer and the connection rules of the nodes between the layers are defined. The nodes within the quay - crane layer are connected based on the adjacent relationship of the quay cranes. The nodes within the transportation layer are connected based on the connectivity of the paths. The nodes within the yard layer are connected based on the adjacency relationship of the yard areas. The inter - layer connection relationships are determined by analyzing the reachability of the operation process. The connection between the quay - crane nodes and the transportation nodes is based on the reachable path of the AGV starting from the quay - crane position. The connection between the transportation nodes and the yard nodes is based on the feasible path of the AGV reaching the yard position. The hierarchical network structure is represented by a graph data structure. The node set contains all the nodes of the three layers, and the edge set contains all the intra - layer and inter - layer connection relationships. Each node in the network structure carries attribute information such as position coordinates, capacity limits, and operation times.

[0042] Construct a connection weight matrix based on the node attribute information in the hierarchical network structure. The weight matrix is a two-dimensional array structure, where the rows and columns correspond to the nodes in the network respectively, and the matrix elements represent the connection weight values between the corresponding nodes. The weight value reflects the cost or priority of the connection between nodes. The distance weight is calculated using the Euclidean distance formula to calculate the spatial distance between nodes. The distance from the quay crane loading and unloading node to the AGV transportation node is calculated by the straight-line distance between two points' coordinates, and the distance from the AGV transportation node to the yard storage node is also determined by coordinate calculation. The shorter the distance, the greater the weight value, indicating a higher connection priority. The physical constraints and operation limitations of the nodes are considered during the weight calculation. The weight calculation of the quay crane node includes the operation capacity and current load condition of the quay crane, the weight calculation of the AGV transportation node includes the passing capacity and traffic condition of the path, and the weight calculation of the yard node includes the storage capacity and current occupancy. The node connection weight table organizes all the calculated weight values according to the node correspondence. The tabular form is convenient for querying and modifying the weight information, and the update mechanism of the weight table dynamically adjusts the weight value according to the real-time operation status of the terminal.

[0043] Associate the node connection weight table with the container transportation capacity constraints. The capacity constraints include the load capacity limit of the AGV and the traffic capacity limit of the path. The association analysis identifies the dependency relationship between the weight value and the capacity constraints, and adjusts the weight value accordingly when the capacity constraints change. The Port-MTCSA algorithm assigns different weights to the transportation paths of 20-foot and 40-foot containers. The differential assignment is based on the weight difference and transportation difficulty difference between the two types of containers. The 20-foot container is lighter and relatively easier to transport, so the corresponding path weight is set to a higher value. The 40-foot container is heavier and more difficult to transport, so the corresponding path weight is set to a lower value. The differential processing is achieved by multiplying the basic weight value by the container type coefficient. The type coefficient of the 20-foot container is greater than 1, and the type coefficient of the 40-foot container is less than 1. The specific value of the coefficient is determined according to the weight ratio and transportation cost ratio of the two types of containers. The capacity constraint weight matrix incorporates the influencing factors of capacity constraints and container types on the basis of the original weight matrix. The calculation of matrix elements comprehensively considers distance weights, capacity weights, and type weights. The dimension of the matrix corresponds to the number of nodes, and the symmetry of the matrix depends on the directionality of the connections between nodes. Based on the capacity constraint weight matrix, the three-layer network structure is finally constructed. During the network construction process, the weight information is integrated into the edge attributes of the network. The weight value of each edge is obtained from the capacity constraint weight matrix. After the network construction is completed, a terminal operation network model is formed. The graph theory algorithm verifies the connectivity of the network. The connectivity verification checks whether there is a reachable path between any two nodes in the network. The verification algorithm uses depth-first search or breadth-first search to traverse all nodes and edges to confirm the integrity and consistency of the network. The verification process includes strong connectivity check and weak connectivity check. Strong connectivity requires that there is a two-way reachable path between any two nodes. Weak connectivity requires that there is a connected path between any two nodes after considering the directed graph as an undirected graph. Select the appropriate connectivity requirement according to the characteristics of the terminal operation process. The terminal operation network topology graph is the final network structure that passes the verification. The graph includes the spatial layout and connection relationships of all operation nodes. The node attributes include information such as location coordinates, operation capabilities, and capacity limits. The edge attributes include information such as distance weights, capacity weights, and type weights. The topology graph provides a network basis for subsequent path planning and task allocation.

[0044] In a specific embodiment, the process of executing step S104 may specifically include the following steps: Collect the terminal historical operation data according to the time node information in the loading-unloading-transportation-storage full-process task chain, and perform data extraction processing on the ship arrival time, operation completion time, and AGV operation trajectory within the past 90 days to obtain the historical operation data set; Arrange the historical operation dataset in a sequence according to the time dimension, and perform pattern recognition processing on the periodic laws and seasonal variations of the ship arrival plan through time series analysis to obtain the ship arrival time series pattern; Collect weather forecast data and historical weather records based on the ship arrival time series pattern, and perform correlation analysis on the correlation between wind speed, rainfall, visibility and terminal operation efficiency to obtain the weather impact coefficient matrix; Extract the type distribution information according to the container transportation records in the historical operation dataset, and perform statistical calculation on the quantity distribution of 20-foot containers and 40-foot containers in different time periods to obtain the container type distribution law; Integrate the ship arrival time series pattern, the weather impact coefficient matrix and the container type distribution law for comprehensive modeling, and perform prediction calculation on the task density of each operation area within the future time window through the load prediction algorithm to obtain the terminal operation load distribution.

[0045] Specifically, collect the terminal historical operation data according to the time node information in the loading-unloading-transportation-storage full-process task chain. The time node information includes the start time, end time and duration of each operation link. The data collection process queries all operation records within the past 90 days from the terminal operation management database, and the query conditions are based on the time stamp field for range screening to ensure the integrity and timeliness of the data. The data extraction process extracts information for three key dimensions: ship arrival time, operation completion time and AGV operation trajectory. The ship arrival time obtains the actual arrival time stamp of each ship from the port scheduling record, the operation completion time extracts the completion time of each container task from the operation logs of quay cranes and yards, and the AGV operation trajectory obtains the movement path and time information of each AGV from the historical records of the vehicle-mounted GPS system. The historical operation dataset is organized in the form of a structured data table, and each row record represents a complete operation event, including fields such as time stamp, operation type, location information, equipment number, task status, etc. The scale of the dataset depends on the operation frequency within 90 days and the detail level of the records.

[0046] Arrange the historical operation dataset in a sequence according to the time dimension. In the sequence arrangement process, all operation records are sorted in the order of timestamps to form a time series data structure. Time series data is a set of data points arranged in chronological order, and each data point contains a time identifier and corresponding numerical or status information. Time series analysis uses statistical methods to identify regular patterns in the data. The analysis process includes three aspects: trend analysis, periodic analysis, and seasonal analysis. Trend analysis identifies the long-term change direction of the data, periodic analysis identifies the repetitive patterns in the data, and seasonal analysis identifies the periodic fluctuation characteristics of the data. Pattern recognition processing detects the regularity of the ship arrival time through autocorrelation analysis and spectral analysis methods. Autocorrelation analysis calculates the correlation of the time series at different time intervals, and spectral analysis identifies the periodic components in the data through Fourier transform. The ship arrival time series pattern includes characteristic parameters such as cycle length, fluctuation amplitude, and peak time, and these parameters reflect the time pattern and change trend of ship arrival.

[0047] Collect weather forecast data and historical weather records based on the ship arrival time series pattern. Weather data collection obtains daily weather information for the past 90 days and weather forecasts for the next week through a meteorological data interface. Weather information includes meteorological parameters such as wind speed, rainfall, visibility, temperature, and humidity. Correlation analysis processing uses the Pearson correlation coefficient calculation method to analyze the linear relationship between weather factors and terminal operation efficiency. The calculation of the correlation coefficient is based on paired samples of historical weather data and operation efficiency data. In the calculation process, weather parameters are used as independent variables, and operation efficiency indicators are used as dependent variables, and the correlation coefficient value is obtained through statistical calculation. Relevance analysis identifies which weather factors have the most significant impact on operation efficiency, and the degree of influence is judged by the absolute value of the correlation coefficient. A correlation coefficient close to 1 indicates a strong positive correlation, close to -1 indicates a strong negative correlation, and close to 0 indicates no correlation. The weather impact coefficient matrix quantitatively represents the impact degree of various weather factors on different operation links. The rows of the matrix represent the types of weather factors, the columns represent the types of operation links, and the matrix elements are the corresponding impact coefficient values. The coefficient values reflect the change range of operation efficiency under specific weather conditions.

[0048] Extract the type distribution information from the container transportation records in the historical operation dataset. The type distribution information extraction process identifies the specification types of each container from the operation records, including two main categories: 20-foot containers and 40-foot containers. The extraction algorithm traverses all transportation records and counts the number of containers of different types. The statistical calculation process groups and statistics the extracted type information according to time periods. The time periods are divided into different granularities such as hours, days, and weeks. The statistical content includes indicators such as the number, proportion, and change trend of containers of each type within each time period. The quantity distribution calculation statistically counts the absolute numbers of 20-foot and 40-foot containers within each time period by the counting method. The proportion distribution calculation obtains the percentage of each type of container in the total quantity through division operations. The change trend calculation identifies the increasing or decreasing trend by comparing the quantity differences in adjacent time periods. The container type distribution law includes time distribution characteristics, quantity proportion characteristics, and change cycle characteristics. The time distribution characteristics reflect the distribution patterns of different types of containers within a day or a week. The quantity proportion characteristics reflect the relative quantity relationship between the two types of containers. The change cycle characteristics reflect the periodic change law of the container type distribution. Integrate the ship arrival time series pattern, the weather impact coefficient matrix, and the container type distribution law for comprehensive modeling. The comprehensive modeling process integrates and processes three different types of data patterns to establish a prediction model affected by multiple factors. The load prediction algorithm is a prediction method based on historical data and influencing factors. The algorithm inputs include time series pattern parameters, weather impact coefficients, and type distribution law parameters. The algorithm output is the load prediction value within the future time window. The prediction calculation process uses the weighted summation method to linearly combine the time series prediction value, the weather impact correction value, and the type distribution prediction value. The weight coefficients are set according to the historical prediction accuracy of each factor. The calculation formula is that the predicted load is equal to the time series prediction value multiplied by the time weight plus the weather correction value multiplied by the weather weight plus the type prediction value multiplied by the type weight. The task density prediction is calculated separately for different operation areas of the terminal. The prediction value of each area is independently calculated based on the historical load data and influencing factors of that area. The terminal operation load distribution includes the expected task density values of each operation area within the future time window. The load distribution data provides load prediction information for the AGV intelligent scheduling decision, solving the technical problem of the lack of predictive scheduling ability in the prior art.

[0049] In a specific embodiment, the process of executing step S105 may specifically include the following steps: Compare each operation area load value in the terminal operation load distribution with a preset load threshold one by one, and perform marked identification processing on the areas exceeding the threshold to obtain a list of overloaded operation areas; Based on the regional location information and the degree of load exceeding in the overloaded operation area list, analyze and evaluate the allocation status of the current AGV resources in each area to obtain the AGV resource reallocation requirement matrix; Call the dynamic replanning algorithm according to the AGV resource reallocation requirement matrix, and perform optimization calculation processing on the transfer path and transfer quantity of AGVs from low-load areas to high-load areas to obtain the AGV resource adjustment plan; Convert the AGV resource adjustment plan into a scheduling instruction format, and reissue the task assignment, path planning, and operation timing of each AGV through the terminal communication network to obtain the optimized scheduling instruction and execute the intelligent scheduling of the terminal AGV.

[0050] Specifically, compare each operation area load value in the terminal operation load distribution with the preset load threshold one by one. The load value comparison process uses a numerical comparison algorithm to judge the size of the current load value of each operation area with the corresponding threshold. The comparison operation is carried out by traversing each element in the load distribution array for one-by-one inspection. When the load value is greater than the threshold, an overloading mark is triggered. The preset load threshold is a critical value determined according to the design capacity of the terminal operation area and historical operation data. The threshold setting takes into account factors such as the operation capacity of quay cranes, the storage capacity of the yard, and the passing capacity of AGVs. Different types of operation areas correspond to different threshold parameters. The threshold for the quay crane area is set based on the maximum operation efficiency of the quay crane, the threshold for the yard area is set based on the storage capacity and stack height limit, and the threshold for the transportation channel is set based on the passing density and traffic flow of AGVs. The mark recognition process performs a marking operation on the areas where the comparison result is true. The mark information includes attributes such as area number, exceeding range, and exceeding time. The overloaded operation area list stores all the marked area information using a list data structure. Each entry in the list contains the area identifier, the degree of load exceeding, and the emergency level. The dynamic update mechanism of the list adds or removes area entries according to the real-time load change situation.

[0051] Based on the regional location information and the degree of load exceeding in the overloaded operation area list, the analysis and evaluation process first extracts the geographical location coordinates and the load exceeding value of each overloaded area from the list. The location information is used to calculate the distance relationship and connectivity between regions, and the degree of load exceeding is used to determine the priority and the quantity of resource allocation. The analysis of the current AGV resource allocation status obtains the real-time location, task status, and load condition of each AGV by querying the AGV management database. The analysis algorithm counts the distribution of the number of AGVs, the number of idle AGVs, and the number of loaded AGVs in each operation area, and calculates the AGV resource density and utilization rate of each area. The evaluation process adopts the load balancing analysis method to calculate the load difference and the unevenness of AGV allocation between regions. The load difference quantifies the dispersion degree of the load distribution between regions through the standard deviation calculation method, and the unevenness of AGV allocation is evaluated by comparing the deviation of the AGV density in each region from the ideal density. The AGV resource reallocation demand matrix uses a two-dimensional matrix structure to represent the resource allocation demands between regions. The rows of the matrix represent the source regions, the columns represent the target regions, and the matrix element values represent the quantity demands of AGVs allocated from the source regions to the target regions. The quantity demands are determined based on the load gap in the target region and the resource redundancy in the source region.

[0052] Call the dynamic re-planning algorithm according to the AGV resource reallocation demand matrix. The dynamic re-planning algorithm is an optimization algorithm that solves complex resource allocation problems by decomposing them into multiple sub-problems. The core idea of the algorithm is to construct the optimal solution of the original problem using the optimal solutions of the sub-problems. The algorithm inputs include the demand matrix, the current status information of AGVs, and the dock path network data. The algorithm output is the optimal AGV allocation plan, including the allocation path and the allocation time sequence of each AGV. The optimization calculation process uses the shortest path algorithm to calculate the optimal transfer path of AGVs from low-load regions to high-load regions. The path optimization considers factors such as distance cost, time cost, and traffic congestion. The optimization of the transfer quantity determines the optimal number of AGVs in each allocation direction through integer programming methods to ensure that the load distribution in each region tends to be balanced after allocation. The path calculation uses the Dijkstra algorithm or the A* algorithm to search for the shortest path from the source node to the target node. The algorithm maintains a distance table to record the shortest distances to each node and continuously updates the distance values through relaxation operations until the optimal path is found. The AGV resource adjustment plan contains the specific allocation instructions for each allocated AGV. The instruction content includes the target region, the transfer path, the estimated arrival time, and the new task assignment. The feasibility verification of the plan checks the traffic capacity and the time window constraints of the allocation path.

[0053] Convert the AGV resource adjustment plan into a scheduling instruction format. During the instruction format conversion process, the abstract allocation information in the adjustment plan is converted into specific operation instructions that the AGV can execute. The conversion process includes path coordinate conversion, speed parameter setting, and task sequence arrangement. The scheduling instruction format adopts a standardized data structure, including fields such as instruction type, target location, execution time, and priority. The instruction encoding adopts the format standard specified by the dock communication protocol to ensure the accurate transmission and parsing of instructions. The dock communication network is a data transmission network connecting the central scheduling center and each AGV. The communication network uses wireless communication technology, supports real-time data transmission and two-way communication, and the network protocol adopts TCP / IP or a dedicated industrial communication protocol to ensure the reliability and real-time nature of data transmission. The reissuance process sends new scheduling instructions to each AGV through the communication network. The issuance process includes steps such as instruction packaging, network transmission, and instruction reception confirmation. The instruction issuance adopts broadcast or point-to-point communication methods according to the urgency and coverage of the instructions. Task allocation reissuance assigns new container transportation tasks to the adjusted AGVs. Path planning reissuance calculates new driving paths and navigation instructions for each AGV. Job timing reissuance adjusts the start time and completion time of each AGV's operation to ensure coordinated operation after deployment. The optimized scheduling instructions are the final instruction set optimized by the dynamic replanning algorithm. The instruction set contains the deployment instructions and new task instructions for all relevant AGVs. The monitoring mechanism for instruction execution tracks the execution status and completion of each AGV.

[0054] The automated terminal AGV intelligent scheduling method in the embodiment of the present application was described above. Next, the automated terminal AGV intelligent scheduling system in the embodiment of the present application will be described. Please refer to Figure 2 , an embodiment of the automated terminal AGV intelligent scheduling system in the embodiment of the present application includes: An evaluation module for collecting container task data through the terminal operation management system, and performing multi-dimensional evaluation processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix; A prediction module for constructing an AGV state vector according to the task priority evaluation matrix, and performing load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with port historical operation data to obtain an AGV load prediction result; A matching module for matching the AGV load prediction result with the task priority evaluation matrix, and performing intelligent matching processing on the container task priority matrix and the AGV dynamic load status vector through the dock multi-task chain scheduling algorithm Port-MTCSA to obtain a full-process task chain for loading, transporting, and stacking; The acquisition module is used to collect the historical operation data of the terminal according to the full process task chain of loading-unloading-transportation-storage, and perform load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the terminal operation load distribution. The comparison module is used to compare the terminal operation load distribution with a preset load threshold. When the predicted load exceeds the threshold, it performs dynamic re-planning processing on the AGV resource allocation, obtains an optimized scheduling instruction, and executes the intelligent scheduling of the terminal AGV.

[0055] above Figure 2 The intelligent scheduling system of the automated terminal AGV in the embodiment of the present invention is described in detail from the perspective of modular functional entities. Next, the intelligent scheduling device of the automated terminal AGV in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0056] Refer to Figure 3 , and an intelligent scheduling device for an automated terminal AGV is further provided in the embodiment of the present invention. The intelligent scheduling device for the automated terminal AGV can be a server, and its internal structure can be as Figure 3 shown. The intelligent scheduling device for the automated terminal AGV includes a processor, a memory, a display screen, an input device, a network interface, and a database connected through a system bus. Among them, the processor of this computer design is used to provide computing and control capabilities. The memory of the intelligent scheduling device for the automated terminal AGV includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the intelligent scheduling device for the automated terminal AGV is used to store the corresponding data in this embodiment. The network interface of the intelligent scheduling device for the automated terminal AGV is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0057] Those skilled in the art can understand that Figure 3 the structure shown in

[0058] is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the intelligent scheduling device of the automated terminal AGV to which the solution of the present invention is applied.

[0059] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, systems, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0060] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this 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 an automated terminal AGV intelligent scheduling device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0061] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An intelligent scheduling method for AGV in an automated terminal, characterized in that, The method includes: Collecting container task data through the terminal operation management system, and performing multi-dimensional evaluation and processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix; Constructing an AGV state vector according to the task priority evaluation matrix, and performing load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with port historical operation data to obtain an AGV load prediction result; Matching the AGV load prediction result with the task priority evaluation matrix, and performing intelligent matching processing on the container task priority matrix and the AGV dynamic load status vector through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a loading-unloading-transportation-storage full-process task chain; Collecting port historical operation data according to the loading-unloading-transportation-storage full-process task chain, and performing load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the port operation load distribution; Comparing the port operation load distribution with a preset load threshold, and when the predicted load exceeds the threshold, performing dynamic re-planning processing on the AGV resource allocation to obtain an optimized scheduling instruction and execute the intelligent scheduling of the terminal AGV.

2. The automated terminal AGV intelligent scheduling method according to claim 1, characterized in that The collecting container task data through the terminal operation management system, and performing multi-dimensional evaluation and processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix includes: Collecting the container number, ship information, quay crane operation plan, yard location, and estimated operation time through the data interface of the terminal operation management system, and performing structured parsing processing on the container task data to obtain a standardized task data set; Calculating the difference between the ship departure time in the standardized task data set and the current time, and performing quantization assignment processing on the urgency coefficient based on the urgency degree of the ship departure time to obtain an urgency score vector; Extracting the weight parameter according to the container specification information in the standardized task data set, and performing classification coding processing on the load levels of 20-foot containers and 40-foot containers to obtain a weight grade coefficient matrix; Performing weighted fusion on the urgency score vector and the weight grade coefficient matrix, and performing comprehensive calculation processing on the quay crane operation sequence, yard distance coefficient, and customer priority through a multi-dimensional weight allocation algorithm to obtain a task priority evaluation matrix.

3. The automated terminal AGV intelligent scheduling method according to claim 1, characterized in that, The constructing an AGV state vector according to the task priority evaluation matrix, and performing load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with port historical operation data to obtain an AGV load prediction result includes: Collecting the real-time operation parameters of the AGV according to the task distribution in the task priority evaluation matrix, and performing vectorization coding processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV to obtain an AGV state vector; Associate the AGV status vector with the terminal historical operation database, and perform feature extraction processing on the AGV load change data within the past 30 days through time window screening to obtain a historical load feature sample set; Based on the historical load feature sample set, perform statistical analysis on the load distribution of AGVs in different yard areas, and quantify the load probabilities of 20-foot and 40-foot containers through load frequency calculation to obtain a load probability distribution table; Perform matching calculation according to the current load status in the AGV status vector and the load probability distribution table, and perform difference analysis on the remaining load capacity of the AGV and the weight requirement of the task to be executed to obtain a load matching degree coefficient; Perform an association operation between the load matching degree coefficient and the expected completion time of the AGV, and predict the change trend of the load capacity of the AGV within the future time window through load time series analysis to obtain the AGV load prediction result.

4. The automated terminal AGV intelligent scheduling method according to claim 1, wherein Match the AGV load prediction result with the task priority evaluation matrix, and perform intelligent matching on the container task priority matrix and the AGV dynamic load status vector through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a full-process task chain for loading, unloading, transportation, and storage, including: Perform data fusion on the AGV load prediction result and the task priority evaluation matrix, and perform matching degree calculation on the container task weight requirement and the AGV load capacity to obtain an initial task-AGV matching table; Based on the initial task-AGV matching table, construct a bipartite graph structure, and perform three-layer network modeling on the quay crane loading and unloading nodes, AGV transportation nodes, and yard storage nodes through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a terminal operation network topology diagram; Extract the task chain path according to the node connection relationship in the terminal operation network topology diagram, and perform chained arrangement on the loading, unloading, transportation, and storage operation sequence of the container from the quay crane to the yard to obtain a candidate task chain set; Perform cumulative cost calculation on the time constraint and load constraint in the candidate task chain set, and minimize the path cost of the task chain from the starting state to the ending state through dynamic programming recursion to obtain a task chain optimization score; Sort and screen the candidate task chains based on the task chain optimization score, and preferentially allocate the high-score task chains through the greedy selection strategy to obtain the matching result of the container task priority matrix and the AGV dynamic load status vector; Verify the matching result of the container task priority matrix and the AGV dynamic load status vector, and perform conflict detection on the load conflict and time conflict in the task chain to obtain a full-process task chain for loading, unloading, transportation, and storage.

5. The automated terminal AGV intelligent scheduling method according to claim 4, wherein, Based on the initial task-AGV matching table, construct a bipartite graph structure, and perform three-layer network modeling on the quay crane loading and unloading nodes, AGV transportation nodes, and yard storage nodes through the terminal multi-task chain scheduling algorithm Port-MTCSA to obtain a terminal operation network topology diagram, including: Extract node information based on the task assignment relationship in the task-AGV initial matching table, perform spatial coordinate mapping processing on the quay crane loading and unloading positions, AGV transportation paths, and yard storage positions to obtain a three-layer node coordinate set; Divide the three-layer node coordinate set according to the quay operation process, and perform topological construction processing on the node connection relationships of the quay crane layer, transportation layer, and yard layer through the quay multi-task chain scheduling algorithm Port-MTCSA to obtain a hierarchical network structure; Construct a connection weight matrix according to the node attribute information in the hierarchical network structure, and perform quantization calculation processing on the distance weight from the quay crane loading and unloading node to the AGV transportation node and the distance weight from the AGV transportation node to the yard storage node to obtain a node connection weight table; Perform correlation analysis on the node connection weight table and the container transportation capacity constraint, and perform differential assignment processing on the transportation path weights of 20-foot and 40-foot containers through the Port-MTCSA algorithm to obtain a capacity constraint weight matrix; Based on the capacity constraint weight matrix, perform the final construction of the three-layer network structure, and verify the connectivity of the quay crane loading and unloading nodes, AGV transportation nodes, and yard storage nodes through graph theory algorithms to obtain a quay operation network topology diagram.

6. The intelligent scheduling method for AGV in an automated terminal according to claim 1, wherein Collect the quay historical operation data according to the loading-unloading-transportation-yard storage full-process task chain, and perform load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the quay operation load distribution, including: Collect the quay historical operation data according to the time node information in the loading-unloading-transportation-yard storage full-process task chain, and perform data extraction processing on the ship arrival time, operation completion time, and AGV operation trajectory within the past 90 days to obtain a historical operation data set; Arrange the historical operation data set in a sequence according to the time dimension, and perform pattern recognition processing on the periodic law and seasonal changes of the ship arrival plan through time series analysis to obtain a ship arrival time series pattern; Collect weather forecast data and historical weather records based on the ship arrival time series pattern, and perform correlation analysis on the correlation between wind speed, rainfall, visibility and quay operation efficiency to obtain a weather impact coefficient matrix; Extract the type distribution information according to the container transportation records in the historical operation data set, and perform statistical calculation processing on the quantity distribution of 20-foot containers and 40-foot containers in different time periods to obtain the container type distribution law; Perform comprehensive modeling on the ship arrival time series pattern, the weather impact coefficient matrix, and the container type distribution law, and perform prediction calculation processing on the task density of each operation area within the future time window through the load prediction algorithm to obtain the quay operation load distribution.

7. The intelligent scheduling method for AGV in an automated terminal according to claim 1, wherein Compare the quay operation load distribution with a preset load threshold, and when the predicted load exceeds the threshold, perform dynamic re-planning processing on the AGV resource allocation to obtain an optimized scheduling instruction and execute the quay AGV intelligent scheduling, including: Compare each operation area load value in the quay operation load distribution with a preset load threshold one by one, and perform marking and identification processing on the areas exceeding the threshold to obtain a list of overloaded operation areas; Based on the area location information and load exceeding degree in the list of overloaded operation areas, analyze and evaluate the allocation status of the current AGV resources in each area to obtain an AGV resource reallocation requirement matrix; According to the AGV resource reallocation requirement matrix, call the dynamic replanning algorithm to optimize the calculation of the transfer path and transfer quantity of the AGV from the low-load area to the high-load area to obtain an AGV resource adjustment plan; Convert the AGV resource adjustment plan into a scheduling instruction format, and reissue the task assignment, path planning, and operation timing of each AGV through the quay communication network to obtain an optimized scheduling instruction and execute the intelligent scheduling of quay AGVs.

8. An automated terminal AGV intelligent scheduling system, characterized in that, For implementing the intelligent scheduling method of quay AGVs as described in any one of claims 1-7, the intelligent scheduling system of quay AGVs includes: An evaluation module for collecting container task data through the quay operation management system, and performing multi-dimensional evaluation processing on the container task data according to the ship departure urgency, container weight grade, quay crane operation sequence, yard distance coefficient, and customer priority to obtain a task priority evaluation matrix; A prediction module for constructing an AGV state vector according to the task priority evaluation matrix, and performing load capacity prediction processing on the current position coordinates, load status, remaining power, and estimated completion time of the AGV through training with the port historical operation data to obtain an AGV load prediction result; A matching module for matching the AGV load prediction result with the task priority evaluation matrix, and performing intelligent matching processing on the container task priority matrix and the AGV dynamic load status vector through the quay multi-task chain scheduling algorithm Port-MTCSA to obtain a full-process task chain of loading-unloading-transportation-storage; A collection module for collecting the quay historical operation data according to the full-process task chain of loading-unloading-transportation-storage, and performing load prediction processing on the ship arrival plan, weather factors, and container type distribution through time series analysis to obtain the quay operation load distribution; A comparison module for comparing the quay operation load distribution with a preset load threshold, and performing dynamic replanning processing on the AGV resource allocation when the predicted load exceeds the threshold to obtain an optimized scheduling instruction and execute the intelligent scheduling of quay AGVs.

9. An AGV intelligent scheduling device for an automated terminal, characterized in that, It includes a memory and a processor. The memory stores a computer program that can run on the processor, and when the processor executes the computer program, it implements the intelligent scheduling method of quay AGVs as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, the processor is caused to execute the intelligent scheduling method of quay AGVs as described in any one of claims 1 to 7.

Images