A port tallying operation management method and system

Abstract

The present application discloses a port tallying operation management method and system, which includes: dividing port tallying operation scenarios, defining and quantifying the node characteristics of each operation node in each operation scenario; constructing a dynamic monitoring strategy model, inputting the operation node characteristics into the dynamic monitoring strategy model, and obtaining a monitoring strategy that matches the operation node; constructing a port tallying operation directed graph, and training and generating a node completion time window prediction model based on the constructed port tallying operation directed graph; obtaining the current operation data of each operation node according to the dynamic monitoring strategy, inputting the node completion time window prediction model, and obtaining the predicted completion time of each operation node; comparing the predicted completion time of each operation node with the preset time of each corresponding operation node, and issuing an operation node progress warning based on the comparison result. The present application can realize accurate intelligent monitoring and warning of port tallying operations, and effectively improve management quality.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent port operation management, and specifically to a port tallying operation management method and system. Background Art

[0002] Logistics management plays a crucial role in the modern economy. Port tallying, a crucial link in the logistics chain, not only ensures the accurate recording of cargo information and the effective guarantee of transportation safety, but also has a direct and far-reaching impact on the efficient and smooth operation of the entire logistics system. Efficient tallying can significantly accelerate cargo turnover, significantly reduce logistics costs, and effectively promote the vigorous development of trade.

[0003] Traditional port tallying operations have long relied heavily on manual record-keeping and monitoring. In the early days, paper records dominated, forcing tally personnel to manually fill out numerous documents to record various aspects of cargo, including type, quantity, and status. This approach was not only cumbersome and inefficient, but also prone to errors due to human negligence, leading to data discrepancies. Furthermore, data collation and querying based on paper records was extremely inconvenient and time-consuming. Advances in technology have led to the emergence of rudimentary digital systems, but these simply converted paper records into electronic documents and failed to fundamentally address the challenges of traditional operations. They generally lacked real-time data transmission and risk warning capabilities, making it impossible to accurately capture and provide feedback on actual on-site conditions during the tallying process. Monitoring data often exhibited lags and errors, failing to truly reflect the real-time status of cargo and the progress of operations. Progress monitoring relied solely on manual estimates and rough statistics, lacking a scientific and precise basis, making accurate monitoring of progress extremely difficult.

[0004] Existing port tally management methods suffer from significant flaws. Due to a lack of timely and reliable data collection and transmission technology and risk warning capabilities, managers struggle to obtain accurate and reliable operational information, hindering effective supervision and management of operational processes. Inaccurate progress monitoring makes it difficult to promptly identify and correct issues in data recording and transmission, resulting in low-quality operational management. Achieving precise intelligent monitoring and early warning of port tally operations and effectively improving management quality are pressing technical challenges in existing technologies. Summary of the Invention

[0005] In order to achieve accurate intelligent monitoring and early warning of port tallying operations and effectively improve management quality, this application provides a port tallying operation management method and system.

[0006] In a first aspect, the present application provides a port tally operation management method, comprising:

[0007] Divide the port tallying operation scenarios, define and quantify the node characteristics of each operation node in each operation scenario;

[0008] Construct a dynamic monitoring strategy model; input the operation node characteristics determined for each operation node in the current operation scenario into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current operation node, including: monitoring resource equipment, monitoring method, and monitoring frequency; the dynamic monitoring strategy model uses a deep learning algorithm to train and generate historical monitoring strategies based on the node characteristics determined for each operation node in each operation scenario in historical port tallying operation data, and the user's satisfaction with the monitoring strategy adopted for the corresponding operation node is greater than a preset satisfaction level;

[0009] For each port tallying operation scenario, the system uses the collected historical port tallying operation data to perform dependency analysis on the data of each operation node, determine the association between the operation nodes, and construct a port tallying operation directed graph. Based on the constructed port tallying operation directed graph, a node completion time window prediction model is trained and generated. According to the matching monitoring strategy, the real-time operation data of each operation node in the current operation scenario is obtained, and the corresponding trained node completion time window prediction model is input into it to obtain the predicted completion time of each operation node.

[0010] The predicted completion time of each job node in the current job scenario is compared with the preset completion time of the corresponding job node, and a job node progress warning is issued for job nodes that exceed the preset completion time of the corresponding job node.

[0011] By adopting the above scheme, the operation scenarios are divided and the node characteristics are quantified to grasp the operation characteristics of different scenarios. The deep learning algorithm is used to build a dynamic monitoring strategy model and match the appropriate monitoring strategy according to the node characteristics to improve the monitoring pertinence and effectiveness. A directed graph is constructed and a node completion time window prediction model is trained to accurately obtain the node predicted completion time. The predicted completion time is compared with the preset time to issue an operation progress warning, and the operation progress delay is discovered in time to ensure that the operation is carried out as planned, thereby improving the efficiency and quality of port tallying operation management.

[0012] Preferably, the dividing of port tallying operation scenarios and the definition and quantification of node characteristics of each operation node in each operation scenario include:

[0013] Classify port tallying operation scenarios based on port type, operation process type, and cargo type, including standard container loading and unloading scenarios at international or domestic ports, bulk cargo loading and unloading scenarios at international or domestic ports, and hazardous materials loading and unloading scenarios at international or domestic ports; define core operation nodes for different operation scenarios, including counting, inspection, classification, marking, damage inspection, and preparation of tally documents; define and quantify the core node characteristics of each operation node in each operation scenario, including: the number of operation node tasks, the complexity of operation node tasks, and the execution risk of operation node tasks, and obtain the quantitative standard values ​​of each operation node characteristic;

[0014] The method of inputting the job node characteristics determined for each job node in the current job scenario into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current job node also includes: introducing a dynamic weight mechanism into the dynamic monitoring strategy model, setting core node feature weight parameters that match different core job nodes, adding feature weight parameters to each input job node feature, and then obtaining a monitoring strategy that matches the current job node.

[0015] By adopting the above scheme, specific division standards are set to divide port tallying operation scenarios, so that operation management can be tailored to actual conditions; core operation nodes under different operation scenarios are defined and the characteristics of core nodes are clarified, which helps to carry out targeted operation management; quantitative standard values ​​of each node characteristic are obtained to improve the accuracy of operation management; a dynamic weight mechanism is introduced into the dynamic monitoring strategy model, and matching core node characteristic weight parameters are set for different core operation nodes, making the obtained monitoring strategy more reasonable and accurate, thereby improving the effectiveness of port tallying operation management.

[0016] Preferably, it also includes:

[0017] For each job node that exceeds the preset completion time of the corresponding job node, calculate the time deviation rate;

[0018] Based on the calculated time deviation rate, the preset time deviation rate range interval of the corresponding operation scenario is determined, and the warning level is matched according to the determined time deviation rate range interval; wherein, each port tallying operation scenario is set with a preset time deviation rate range interval of the corresponding operation scenario, and each set time deviation rate range interval is preset with a warning level matching it;

[0019] Emergency strategies are preset according to the warning level combination of operation scenarios and operation nodes, including: operation resource allocation strategy, operation process optimization strategy and operation auxiliary technology intervention strategy; for the current port tallying operation scenario, the preset emergency strategies are matched and executed for operation nodes at different warning levels.

[0020] By adopting the above solution, the warning level is accurately matched according to the time deviation rate of the operation node, and corresponding emergency strategies are implemented for operation nodes of different warning levels, such as operation resource allocation, operation process optimization and operation auxiliary technology intervention, effectively ensuring the smooth progress of port tallying operations.

[0021] Preferably, it also includes:

[0022] For each port tallying operation scenario, define the emergency events during the port tallying operation and the operational risk data caused by the emergency events, including: risk type, transmission probability and time delay coefficient;

[0023] Risk transmission edge modeling includes defining edge attributes such as risk type, transmission probability, and time delay coefficient. Risk transmission edges are superimposed on the original directed graph dependency edges to form a dynamic extended graph, enabling updates to the original port tallying operation directed graph.

[0024] Based on the updated directed graph of port tallying operations, a risk-corrected node completion time window prediction model is trained and generated to obtain the risk-corrected node completion time by utilizing the collected historical port tallying operation data, historical operating environment data, historical operating risk data obtained based on the historically collected port tallying operation data and operating environment data, and the historical risk-corrected node completion time corresponding to each operation node.

[0025] By adopting the above scheme, the emergency events and operational risk data during the port tallying operation are defined, so that the system can more comprehensively consider various risk factors that may affect the operation; by modeling the risk transmission edge and updating the port tallying operation directed graph, the risk factors can be incorporated into the original operation process model, making the model more in line with the actual situation; completing multi-task joint training to generate a risk-corrected node completion time window prediction model, more accurately predicting the completion time of the operation node, comprehensively considering the normal operation process and potential risks, thereby improving the accuracy of operation progress monitoring and effectively preventing and responding to potential risks in the tallying process.

[0026] Preferably, emergency strategies that match the combination of operation scenario, risk type and operation node warning level are preset to replace the emergency strategies that match the combination of operation scenario and operation node warning level; and a collective emergency strategy is preset for each combination of operation scenario, risk type and operation node warning level, and the elements of the preset collective emergency strategy are emergency strategies with different response time limits and are respectively provided with a range of comprehensive risk indexes that match them, and the emergency strategy with a lower response time limit has a smaller value of the range of comprehensive risk indexes that match it;

[0027] Using a deep learning algorithm to determine the risk impact and risk response difficulty of each operation node based on the operation data and operation environment data obtained from the early warning, and calculate a comprehensive risk index; the comprehensive risk index is calculated based on the weighted risk impact and risk response difficulty;

[0028] In the current port tallying operation scenario, the preset collective emergency strategy is matched with the operation node based on the obtained risk type and warning level combination; the emergency strategy is continued to be matched according to the comprehensive risk index range interval of the calculated risk comprehensive index.

[0029] By adopting the above solution, emergency strategies that match the operation scenario, risk type, and operation node warning level are preset to respond to different situations more accurately. For a single combination, multiple emergency strategies with different response time limits are preset and corresponding comprehensive risk index ranges are set, so that appropriate strategies can be flexibly selected based on different risks. The comprehensive risk index is calculated and the emergency strategy is matched accordingly, so that emergency measures are more in line with the actual risk situation, further improving the effectiveness and pertinence of responding to potential risks and abnormal situations in tallying operations, and reducing losses caused by risks.

[0030] Preferably, it also includes: providing a job node progress visualization window to display the job progress of each job node in real time; providing a task execution process interface to support users to adjust the job nodes, receiving the job node task execution process adjusted by the user and verifying whether it is inconsistent with the task execution mandatory process of the job node, and generating a job node task execution process adjustment failure prompt when it is verified to be inconsistent; or automatically adjusting the task execution process of each job node according to the progress percentage or expected completion time of each job node, so as to give priority to the task execution process corresponding to the job node with a small progress percentage or a later expected completion time, and there is no contradiction between the adjusted job node execution process and the task execution mandatory process of the job node.

[0031] By adopting the above solution, a visual display of the job progress is achieved, which facilitates real-time understanding of the job status. At the same time, it supports manual and automatic adjustment of the job node execution process. Under the premise of ensuring that the mandatory process is not violated, node tasks with slow progress or long time consumption are executed first, thereby improving the overall efficiency of the job.

[0032] Preferably, the monitoring method includes a centralized monitoring method or a distributed monitoring method; the monitoring resource equipment includes: Internet of Things sensors, 5G real-time transmission equipment, AR auxiliary equipment, AI visual equipment, scanning devices; the monitoring frequency includes: real-time monitoring, periodic monitoring or triggered monitoring.

[0033] By adopting the above solution, the monitoring methods, monitoring resources and equipment, and monitoring frequency are refined, and appropriate monitoring modes and resources are flexibly selected according to different port tallying operation scenarios and operation nodes to meet diverse monitoring needs, thereby monitoring the operation process more accurately and effectively, ensuring accurate control and smooth progress of the operation.

[0034] In a second aspect, the present application provides a port tally operation management system, comprising:

[0035] The port tallying operation definition module is used to divide the port tallying operation scenarios, define and quantify the node characteristics of each operation node in each operation scenario;

[0036] The port tallying operation monitoring strategy acquisition module is used to build a dynamic monitoring strategy model. The operation node characteristics determined for each operation node in the current operation scenario are input into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current operation node, including monitoring resources and equipment, monitoring methods, and monitoring frequency. The dynamic monitoring strategy model uses a deep learning algorithm to train and generate historical monitoring strategies based on the node characteristics determined for each operation node in each operation scenario in historical port tallying operation data, as well as historical monitoring strategies where user satisfaction with the monitoring strategy adopted for the corresponding operation node exceeds a preset satisfaction level.

[0037] The port tallying operation completion time prediction module is used to analyze the dependency relationships of the data of each operation node based on the collected historical port tallying operation data for each port tallying operation scenario, determine the association relationship between the operation nodes, construct a port tallying operation directed graph, and train and generate a node completion time window prediction model based on the constructed port tallying operation directed graph. According to the matching monitoring strategy, the real-time operation data of each operation node in the current operation scenario is obtained, and the corresponding trained node completion time window prediction model is input into it to obtain the predicted completion time of each operation node.

[0038] The port tallying operation progress warning module is used to predict the completion time of each operation node in the current operation scenario, compare it with the preset completion time of the corresponding operation node, and issue an operation node progress warning for the operation node that exceeds the preset completion time of the corresponding operation node.

[0039] By adopting the above scheme, dividing the port tallying operation scenarios and defining and quantifying the node characteristics can make the operation management more targeted; building a dynamic monitoring strategy model and combining the node characteristics to obtain the matching monitoring strategy can achieve accurate monitoring; using historical data to build a directed graph and generate a node completion time window prediction model can predict the completion time of the operation node; issuing early warnings for operation nodes that exceed the preset time, timely discovering operation progress problems, thereby improving the management efficiency and quality of port tallying operations and ensuring the smooth progress of operations.

[0040] In a third aspect, the present application provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method as described above.

[0041] In a fourth aspect, the present application provides a computer device, which includes a memory, a processor, and a program stored and executable on the memory, and the program implements the steps of the above method when executed by the processor.

[0042] In summary, this application has the following beneficial effects:

[0043] 1. By dividing the operation scenarios, defining quantitative node characteristics, and combining them with a dynamic monitoring strategy model to obtain a matching monitoring strategy, timely and reliable collection of real-time operation data is achieved. A node completion time window prediction model is used to obtain the predicted completion time of each node. This enables real-time monitoring and progress warnings of tallying operations, improves the real-time nature of data recording and transmission, and solves the problem of accurate operation progress monitoring.

[0044] 2. By leveraging a dynamic monitoring strategy model that employs deep learning algorithms and historical data training, as well as dependency analysis and risk assessment prediction of operation node data, it can effectively predict potential risks in the tallying process and obtain corresponding corrections to the time required to complete tallying operations under specific risk types, enabling early warning and intervention, and reducing the probability of accidents.

[0045] 3. Provide a visualization window for the progress of job nodes and a task execution process interface that supports user adjustment of job nodes, or automatically adjust the execution process according to the progress of job nodes, to improve the flexibility and response speed of on-site operations and facilitate the adjustment of work processes in response to emergencies. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of the port tally operation management method described in a specific embodiment;

[0047] Figure 2 It is a structural diagram of the port tally operation management system described in a specific embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0049] like Figure 1As shown, an embodiment of the present application discloses a port tallying operation management method, which achieves the effect of improving the real-time, accuracy and flexibility of port tallying operation management by adopting steps including operation scene division, monitoring strategy matching, correlation analysis, time prediction, progress warning, emergency strategy formulation, visualization display and process adjustment. The present application is further described in detail below.

[0050] S1. Divide the port tallying operation scenarios, define and quantify the node characteristics of each operation node in each operation scenario.

[0051] Taking into account the differences in operation nodes in different port tallying operation scenarios, such as the more complicated inspection process for dangerous goods; in order to more accurately monitor the progress of each operation node, appropriate intelligent monitoring strategies are selected to timely collect real-time data of each operation node to assist in subsequent progress monitoring and early warning of each operation node.

[0052] In this embodiment, the operation scenario classification includes classification based on port type (such as international port or domestic port), operation process type (such as warehousing scenario and outbound scenario), and cargo type (such as standard containers with uniform specifications and conventional goods, irregularly shaped bulk cargo and conventional goods, and dangerous goods with high safety risks). The corresponding combination-generated scenario types include: international port standard container outbound scenario, international port standard container inbound scenario, domestic port standard container inbound scenario, domestic port standard container outbound scenario, international port bulk cargo outbound scenario, international port bulk cargo inbound scenario, domestic port bulk cargo inbound scenario, domestic port bulk cargo outbound scenario, international port dangerous goods inbound scenario, international port dangerous goods outbound scenario, domestic port dangerous goods inbound scenario, and domestic port dangerous goods outbound scenario, etc.

[0053] At the same time, the core operation nodes under different operation scenarios are defined. Considering that various tallying operation scenarios involve core operation nodes such as counting, inspection, classification, marking, damage inspection and compilation of tallying documents; however, the core operation nodes corresponding to different tallying operation scenarios are different. According to the characteristic description of each operation node, the core node characteristics of each operation node in each operation scenario are defined and quantified, including: the number of operation node tasks, the complexity of operation node tasks and the execution risk of operation node tasks, and the quantitative standard value of each operation node characteristic is obtained.

[0054] Node feature quantification can be determined by establishing a reasonable scoring system or employing a fuzzy evaluation method. In this embodiment, for each operation scenario, the number of operation node tasks (e.g., a score of 0-1) within the core node characteristics of each operation node can be statistically scored based on the number of cargo units and document processing volume. For example, the number of cargo units is calculated as log(actual volume) / log(baseline volume), with the baseline volume set as: standard container = 100 TEU, breakbulk = 500 tons, and hazardous goods = 30 tons; document processing volume is calculated as the number of document pages / 20 (with a maximum score of 1). For each operation scenario, the complexity of the operation node tasks (e.g., a score of 0-1) within the core node characteristics of each operation node is determined based on the cargo type and operational requirements. The heterogeneity of the cargo (e.g., whether it is hazardous goods or refrigerated goods) and the number of operational steps (e.g., whether customs inspection, sorting, and reorganization are required) are then added to the base score (preset to 0.3). For example, if the cargo type is hazardous goods, 0.3 points are added to the base score; if the cargo requires sorting and reorganization, 0.3 points are added to the base score. For each operation scenario, the core node characteristics of each operation node (e.g., 0-1) can be used to determine the execution risk of the operation node task. This risk can be determined based on the cargo value coefficient, regulatory level, and historical error rate. For example, risk = max(cargo value coefficient, regulatory coefficient) * (1 + historical error rate). The cargo value coefficient can be determined by setting a base cargo value coefficient and obtaining a ratio, such as min(total cargo value / 1.6e, 1.0). The regulatory coefficient can be obtained by matching the regulatory level of different goods, such as general cargo = 0.3, bonded cargo = 0.6, and dangerous goods = 0.9. Accordingly, the quantitative standard values ​​of each operation node characteristic are obtained to generate a scenario-specific node feature matrix, such as [0.9, 0.5, 0.4] for the domestic port bulk cargo outbound counting operation node.

[0055] S2. Build a dynamic monitoring strategy model; input the operation node characteristics determined for each operation node in the current operation scenario into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current operation node.

[0056] Specifically, the dynamic monitoring strategy model adopts a deep learning algorithm, and is generated by training the obtained node features of each operation node in each operation scenario in the historical port tallying operation data, and the historical monitoring strategy in which the user's satisfaction with the monitoring strategy adopted for the corresponding operation node is greater than the preset satisfaction level; wherein, the degree of monitoring satisfaction can be determined based on the satisfactory feedback value of users / experts on the timeliness of collecting data and analyzing the progress of the operation node according to the monitoring strategy under the conditions of the historical adoption of a specific monitoring strategy.

[0057] Focusing on the needs of intelligent port tallying operations and combining the characteristics of different scenarios, AR, 5G, and the Internet of Things (IoT) are integrated to build a more accurate operation progress monitoring strategy to ensure efficient and accurate operation progress. Specifically, the monitoring strategy includes monitoring methods, monitoring resources and equipment, and monitoring frequency, which are combined to generate a corresponding monitoring strategy. Monitoring methods include centralized monitoring (using IoT sensors to rapidly collect information from multiple operation nodes and transmit it to a monitoring center in real time via the 5G network) or distributed monitoring (dividing the entire operation process into multiple sub-nodes, setting up independent monitoring modules for each sub-node, collecting operation status information from each sub-node via IoT devices, and transmitting data via the 5G network). Centralized monitoring can be used for operation nodes with more than the preset number of tasks, while distributed monitoring can be used for operation nodes with less than the preset number of tasks. Furthermore, based on other descriptive characteristics of each operation node (such as whether it has any relationship with other operation nodes), independent monitoring (for relatively independent operation nodes without direct relationships with other operation nodes) or linked monitoring (for operation nodes with direct relationships with other operation nodes, such as linked operation nodes, a combination of monitoring equipment and monitoring frequency can be selected).

[0058] Monitoring resource equipment includes IoT sensors, 5G real-time transmission equipment, AR auxiliary equipment (such as handheld AR devices, AR glasses), AI visual equipment (such as smart cameras, AI visual robots, etc.), scanning devices (such as smart scanners), etc. Different operation nodes can be matched with different monitoring resource equipment combinations. For example, for operation nodes with complexity lower than the preset complexity, conventional monitoring resource equipment can be used. For operation nodes with complexity not lower than the preset complexity, an enhanced monitoring strategy including AR auxiliary equipment and AI visual equipment can be adopted.

[0059] Monitoring frequencies include real-time monitoring, periodic monitoring or triggered monitoring. Different operation nodes can match different monitoring frequencies. For example, different operation nodes can match different monitoring resource equipment combinations. For operation nodes with a risk level lower than the preset risk level, periodic monitoring or triggered monitoring can be used. For operation nodes with a risk level lower than the preset risk level, real-time monitoring can be used.

[0060] Based on the constructed and trained dynamic monitoring strategy model, the node features determined for each node in the current job scenario are input into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current node. Furthermore, considering the different monitoring priorities of different core nodes, namely, the counting node prioritizes the number of task points, a dynamic weighting mechanism is introduced into the dynamic monitoring strategy model to adjust the weights of each node's features to match a more appropriate monitoring strategy. Specifically, this includes setting matching weight parameters for core node features for different core nodes, such as (0.5, 0.3, 0.2) for the counting node, (0.3, 0.4, 0.3) for the inspection node, and (0.2, 0.5, 0.3) for the classification node. Furthermore, feature weight parameters are added to each input node feature to obtain a monitoring strategy that matches the current node.

[0061] S3. Construct a directed graph of port tallying operations, and train and generate a node completion time window prediction model based on the constructed directed graph of port tallying operations.

[0062] Taking into account the dependencies among various operation nodes in some ports, such as the order of counting, inspection, classification, and marking, once the completion time of a certain operation node is delayed, it will accordingly affect the progress of subsequent operation nodes. Considering the relationship between different operation nodes, a directed graph of port tallying operations is established for each port tallying operation scenario, and relevant graph structure training is carried out to better predict the completion time of each operation node. Specifically, for each port tallying operation scenario, a directed graph of port tallying operations is constructed, including: for each port tallying operation scenario, the collected historical port tallying operation data is used to clean and integrate the data, and the dependency analysis is performed on the data of each operation node. The characteristics of the operation nodes are extracted (including: core characteristics of the operation node characteristics, historical average operation node time consumption, operation node type, operation node resource requirements, operation node corresponding cargo attributes and other characteristic data), and the association between the operation nodes is determined (such as: timing dependency, resource competition relationship, etc.). The directed graph of port tallying operations G=(V,E) is constructed, where the operation node V represents the specific task of the operation node (such as "damage inspection" and "cargo classification"), and the edge E represents the dependency relationship (such as "cargo classification by destination after inspection"). The node attributes and edge attributes are converted into low-dimensional vectors through the embedding layer.

[0063] For each port tallying operation scenario, a node completion time window prediction model corresponding to the port tallying operation scenario is trained and generated based on the constructed port tallying operation directed graph. Specifically, each node completion time window prediction model adopts a graph neural network architecture design, using a graph convolutional network to aggregate the features of adjacent nodes and capture spatial features. A long short-term memory network or Transformer encoder is introduced to process the temporal characteristics of the node's historical time consumption and perform spatiotemporal feature fusion. The model input is the real-time operation data of the operation node in a specific scenario and the port tallying operation directed graph. The model output is the predicted earliest completion time (EST) and latest completion time (LST) of each operation node, forming a time window [EST, LST]. The node completion time window prediction model is trained by embedding the real-time data of the historical operation nodes and the completion time of each operation node in the port tallying operation directed graph.

[0064] According to the matching monitoring strategy, the real-time operation data of each operation node in the current operation scenario is obtained, and the corresponding trained node completion time window prediction model is input to obtain the predicted completion time of each operation node.

[0065] S4. Forecast the completion time of each operation node in the current operation scenario, compare it with the preset completion time of the corresponding operation node, and issue an operation node progress warning for the operation node that exceeds the preset completion time of the corresponding operation node.

[0066] Specifically, the predicted completion time of each operation node in the current operation scenario is compared with the preset completion time of the corresponding operation node (that is, the ideal completion time set by the user for the port tallying operation requirements). For operation nodes that exceed the preset completion time of the corresponding operation node, an operation node progress warning is issued to remind the operation personnel in time.

[0067] To provide more intuitive early warnings, the time deviation rate is calculated for each operation node that exceeds the corresponding preset completion time. The formula is: (the predicted latest completion time of the operation node - the preset time of the preset operation node) / the preset completion time of the operation node. Based on the calculated time deviation rate, the preset time deviation rate range of the corresponding operation scenario is determined, and the warning level is matched based on the determined time deviation rate range. Each port tallying operation scenario has a preset time deviation rate range, and each preset range has a matching level.

[0068] It provides a job progress visualization window to display the job progress of each job node in real time and the warning level of each job node in different colors. It can also synchronize the corresponding information of the job progress visualization window to each AR device through 5G transmission technology, so that operators can check the job progress of the current job node in time.

[0069] In addition to the above-mentioned operation progress warning, in order to complete the port tallying operation more promptly, corresponding preset emergency strategies are set for operation nodes that may cause operation delays. The method also includes:

[0070] S5. Preset an emergency strategy that matches the operation scenario and the operation node warning level combination, and match the preset emergency strategy to the operation nodes of different warning levels obtained in each current port tallying operation scenario and execute it.

[0071] Specifically, a preset emergency strategy matching each combination is set in advance based on the combination of the operation scenario and the operation node warning level, wherein the higher the priority in different operation scenarios (the priority of different operation scenarios is pre-set), the higher the corresponding quantitative first value, and the higher the priority of different operation node warning levels (the priority of different operation node warning levels is pre-set), the higher the corresponding quantitative first value. The larger the weighted calculation result of the quantitative first value and the second data, the faster the preset emergency strategy with shorter emergency response time will be matched.

[0072] Among them, the preset emergency strategies generally include: operation resource allocation, operation process optimization and operation auxiliary technology intervention and other strategies. The preset emergency strategies with shorter emergency response time correspond to more operation resource allocation, shorter total operation time after operation process optimization and more advanced operation auxiliary technology intervention. For example, under specific operation scenario conditions, the inspection operation node with a red warning level matches the first preset emergency strategy, including: allocating more manpower and equipment resources to the operation node, and optimizing the operation process (such as: automatically adjusting the task execution process of each operation node according to the progress percentage or expected completion time of each operation node, giving priority to the task execution process corresponding to the operation node with a small progress percentage or a later expected completion time, and there is no conflict between the adjusted operation node execution process and the task execution mandatory process of the operation node), and introducing advanced operation auxiliary technologies (such as handheld AR inspection equipment) to improve operation efficiency.

[0073] For the operation nodes with different warning levels obtained in each current port tallying operation scenario, the preset emergency strategy is matched and executed. In order to further obtain accurate emergency strategies, a deep learning algorithm can be used to construct an emergency strategy acquisition model. The historical emergency strategy training is generated through the warning level data of historical port tallying operation scenarios and operation nodes, and the historical emergency strategy adopted by the corresponding node after the operation node delay rate is greater than the preset delay rate.

[0074] In addition, considering that some users may have the need to independently adjust the task execution of job nodes due to the possibility of job node timeout, a corresponding task execution process interface is provided to support users in adjusting the job nodes. The task execution process of the job node adjusted by the user is received and verified whether it is inconsistent with the mandatory task execution process of the job node. When the verification is inconsistent, a prompt for failure to adjust the task execution process of the job node is generated.

[0075] Using the above method, we determine the node characteristics by dividing the scene, input the model to obtain the monitoring strategy, analyze the node correlation and predict the completion time, issue warnings for timed nodes, formulate emergency strategies, display the operation progress and adjust the process, so as to achieve the effect of real-time and accurate monitoring of operation progress, timely warning of risks, and flexible adjustment of operation processes.

[0076] In a specific embodiment, taking into account the possibility that the operation progress may be stagnant due to environmental or equipment factors, the normal operation process and potential risks are comprehensively considered to improve the accuracy of operation progress monitoring and effectively prevent and deal with potential risks in the tallying process. The method further includes:

[0077] For each port tallying operation scenario, data on emergencies during port tallying operations and the operational risks caused by these emergencies are defined, including risk type, transmission probability, and time delay coefficient. Specifically, for each port tallying operation scenario, a sudden event database is constructed to store sudden events and the operational risk data caused by these sudden events. In this embodiment, sudden events are defined as severe weather, operational accidents, and externally restricted operations. Based on historical port tallying operation data for each operation scenario, statistical analysis is performed to obtain the corresponding risk type, transmission probability, and time delay coefficient for each sudden event. For example, for risk types such as severe weather (typhoons, heavy fog, etc.), operational accidents (equipment failures, cargo collapse, etc.), and externally restricted operations (power restrictions, port restrictions, etc.), the transmission probability is calculated based on historical data to determine the probability of a specific risk type sudden event affecting the associated operation node. For example, the probability of typhoon weather causing inspection delays is 70%. The time delay coefficient is calculated based on the time lag after the risk is transmitted to the associated operation node. For example, the delay caused by severe weather on unloading is 30 minutes (calculated based on historical typhoon operation delay data).

[0078] For each port tallying operation scenario, risk transmission edge modeling is performed, including defining edge attributes including: risk type (marking the risk type of transmission), transmission probability (the probability of a risk event triggering a delay in the associated node) and time delay coefficient (the delay of risk transmission); on the basis of the original directed graph dependency edge, the risk transmission edge is superimposed to form a dynamic extension graph, realizing the update of the original port tallying operation directed graph ,in, Represents the risk transmission edge set.

[0079] For each port tallying operation scenario, based on the corresponding updated port tallying operation directed graph, using the collected historical port tallying operation data, historically collected operation environment data, historical operation risk data obtained based on the historically collected port tallying operation data and operation environment data, and the historical risk-corrected node completion time of each operation node, the multi-task joint training of node completion time window prediction and risk assessment prediction is completed, and a risk-corrected node completion time window prediction model is generated to obtain the risk-corrected node completion time. Among them, for each port tallying operation scenario, the corresponding risk-corrected node completion time window prediction model is designed using a graph neural network architecture. In addition to designing a spatiotemporal convolutional graph, using a graph convolutional network to aggregate the features of adjacent nodes, capturing node spatial features, and using LSTM or Transformer to encode the historical time-consuming sequence of nodes, a risk transmission module is also designed, including a probability propagation layer and a delay adjustment layer. The probability propagation layer uses a graph convolutional network to aggregate the risk transmission probability of neighboring nodes and calculate the comprehensive risk value of the node. For job nodes The overall risk probability, is the edge conduction weight; the delay adjustment layer modifies the timing characteristics based on the delay coefficient of the risk conduction edge to obtain the risk conduction delay The model inputs are real-time operation data for specific scenarios and an updated directed graph of port tallying operations. The model output is node completion times with risk confidence intervals. By embedding historical real-time operation data of operation nodes in the directed graph of port tallying operations, historical operation risk data derived from historically collected port tallying operation and operating environment data, and the risk-adjusted node completion times for each operation node (calculated based on completion times and risk transmission delays), a prediction model for node completion risk-adjusted node completion time windows is trained. For example, if the temperature check operation node is a risk node with a transmission probability of 0.8 and a delay of 1.5 hours, the corresponding "check-mark" risk transmission edge attributes include a transmission probability of 0.8 and a delay of 1.5. The original prediction time window for the marking operation node is [10:00, 12:00], which is adjusted to [11:30, 14:30] after considering the risk, with a confidence level of 80%.

[0080] In addition, due to different risk types, preset emergency strategies can be further set for different risk types to assist in subsequent matching of appropriate preset emergency strategies; the method further includes:

[0081] Emergency strategies that match the combination of operation scenarios, risk types and operation node warning levels are preset to replace the emergency strategies that match the combination of operation scenarios and operation node warning levels; and for each combination of operation scenarios, risk types and operation node warning levels, a set emergency strategy is preset, and the elements in the preset set emergency strategy are emergency strategies with different response time limits and are respectively set with matching comprehensive risk index range intervals. The lower the response time limit, the smaller the value of the comprehensive risk index range interval matched by the emergency strategy; for example: under specific operation scenario conditions, the risk type is typhoon weather, and the warning level is red for the inspection operation node, the first set of preset emergency strategies is matched; the first set of preset emergency strategies (each set of preset emergency strategies) are all sorted according to the response time limit. The lower the response time limit, the higher the priority and the higher the ranking, and each response time limit corresponds to the preset emergency strategy and is set with a threshold-matched comprehensive risk index range interval.

[0082] For the operation data and operation environment data of each operation node for early warning in the corresponding operation scenario obtained, a comprehensive risk index is calculated; the comprehensive risk index is weighted according to the risk impact degree and the difficulty of risk response; wherein, the risk impact degree of the operation node can be determined by analyzing the operation data and the operation environment to determine the transmission probability and the operation node and the associated operation node, and then perform weighted calculation; or it can be determined based on the operation data and operation environment data of the operation node, and specifically for each operation environment, a risk impact degree module can be constructed using neural network data; the module inputs the operation data and operation environment data of the operation node in the current operation scenario, extracts the operation risk data characteristics, and then obtains the risk impact degree value, and through the corresponding operation scenario The operation data and operation environment data of the historical operation nodes and the risk impact degree marked by experts are used for training; accordingly, the risk response difficulty of the operation node is determined according to the operation data and operation environment data of the operation node. Specifically, for each operation environment, a risk response difficulty identification module can be constructed using neural network data; this module inputs the operation data and operation environment data of the operation node in the current operation scenario, extracts the operation risk data features, and then obtains the risk response difficulty value, and trains the risk response difficulty through the operation data and operation environment data of the historical operation nodes in the corresponding operation scenario and the risk response difficulty marked by experts; for example: the operation data and operation environment data of the historical operation nodes corresponding to different levels of typhoon weather are marked with different response difficulty values.

[0083] For the current port tallying operation scenario, the preset set emergency strategy is matched with the operation node based on the obtained risk type and warning level combination, and the emergency strategy is continued to be matched according to the comprehensive risk index range of the calculated risk comprehensive index.

[0084] like Figure 2As shown, this embodiment discloses a port tally operation management system, which specifically includes:

[0085] The port tallying operation definition module 101 is used to divide the port tallying operation scenarios, and define and quantify the node characteristics of each operation node in each operation scenario;

[0086] The port tallying operation monitoring strategy acquisition module 102 is used to build a dynamic monitoring strategy model. The operation node characteristics determined for each operation node in the current operation scenario are input into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current operation node, including monitoring resources and equipment, monitoring methods, and monitoring frequency. The dynamic monitoring strategy model uses a deep learning algorithm to train and generate historical monitoring strategies based on the node characteristics determined for each operation node in each operation scenario in historical port tallying operation data, and the user's satisfaction with the monitoring strategy adopted for the corresponding operation node is greater than a preset satisfaction level.

[0087] The port tallying operation completion time prediction module 103 is used to use the collected historical port tallying operation data for each port tallying operation scenario, perform dependency analysis on the data of each operation node, determine the association relationship between the operation nodes, construct a port tallying operation directed graph, and train and generate a node completion time window prediction model based on the constructed port tallying operation directed graph; obtain the real-time operation data of each operation node in the current operation scenario according to the matching monitoring strategy, input the corresponding trained node completion time window prediction model, and obtain the predicted completion time of each operation node;

[0088] The port tallying operation progress warning module 104 is used to predict the completion time of each operation node in the current operation scenario, compare it with the preset completion time of the corresponding operation node, and issue an operation node progress warning for the operation node that exceeds the preset completion time of the corresponding operation node.

[0089] In a specific embodiment, the system further includes:

[0090] The port tallying operation emergency strategy acquisition module 105 is used to calculate the time deviation rate for each operation node that exceeds the preset completion time of the corresponding operation node; determine the preset time deviation rate range interval of the corresponding operation scenario based on the calculated time deviation rate, and match the warning level based on the determined time deviation rate range interval; wherein, each port tallying operation scenario is set with a preset time deviation rate range interval of the corresponding operation scenario, and each set time deviation rate range interval is preset with a matching warning level; preset emergency strategies are preset according to the combination of the operation scenario and the warning level of the operation node, including: operation resource allocation strategy, operation process optimization strategy and operation auxiliary technology intervention strategy; for the current port tallying operation scenario, the preset emergency strategies are matched and executed for the operation nodes at different warning levels.

[0091] In a specific embodiment, the system further includes:

[0092] The port tallying operation completion time optimization prediction module 106 is also used to define, for each port tallying operation scenario, emergency events during the port tallying operation and operational risk data caused by the emergency events, including: risk type, transmission probability and time delay coefficient; risk transmission edge modeling, including defining edge attributes including: risk type, transmission probability and time delay coefficient; on the basis of the original directed graph dependency relationship edges, superimposing risk transmission edges to form a dynamic extension graph, thereby updating the original constructed port tallying operation directed graph; based on the updated port tallying operation directed graph, using the collected historical port tallying operation data, the historically collected operating environment data, the historical operating risk data obtained based on the historically collected port tallying operation data and the operating environment data, and the historical risk-corrected node completion time corresponding to each operation node, train and generate a risk-corrected node completion time window prediction model to obtain the risk-corrected node completion time.

[0093] The port tallying operation emergency strategy optimization module 107 is also used to preset a matching emergency strategy according to the combination of operation scenario, risk type and operation node warning level to replace the preset matching emergency strategy according to the combination of operation scenario and operation node warning level; and for each combination of operation scenario, risk type and operation node warning level, a collective emergency strategy is preset, and the elements in the preset collective emergency strategy are emergency strategies with different response time limits and are respectively set with matching comprehensive risk index range intervals, and the emergency strategy with a lower response time limit matches a smaller comprehensive risk index range interval value; using a deep learning algorithm to determine the risk impact degree and risk response difficulty of the operation node for the operation data and operation environment data of each operation node obtained by the warning, and calculate the risk comprehensive index; the risk comprehensive index is weightedly calculated according to the risk impact degree and risk response difficulty; in the current port tallying operation scenario, the preset collective emergency strategy is matched with the operation node of the obtained risk type and warning level combination; and the emergency strategy is continued to be matched according to the comprehensive risk index range interval in which the calculated risk comprehensive index is located.

[0094] The embodiment of the present application also discloses a computer-readable storage medium.

[0095] Specifically, the computer-readable storage medium stores a computer program that can be loaded by a processor and executed by the port tallying operation management method described above. The computer-readable storage medium includes, for example, various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0096] The embodiment of the present application also discloses a computer device.

[0097] Specifically, the computer device includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the above-mentioned port tally operation management method.

[0098] The above are all preferred embodiments of the present application and are not intended to limit the scope of protection of this application. Unless otherwise stated, any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features. In other words, unless otherwise stated, each feature is merely an example of a series of equivalent or similar features.

Claims

1. A port tally operation management method, characterized in that: include: Divide the port tallying operation scenarios, define and quantify the node characteristics of each operation node in each operation scenario; Build a dynamic monitoring strategy model; The operation node characteristics determined for each operation node in the current operation scenario are input into a dynamic monitoring strategy model to obtain a monitoring strategy that matches the current operation node, including monitoring resources and equipment, monitoring methods, and monitoring frequency. The dynamic monitoring strategy model uses a deep learning algorithm to train and generate node characteristics determined for each operation node in each operation scenario from historical port tallying operation data, as well as historical monitoring strategies where user satisfaction with the monitoring strategy adopted for the corresponding operation node exceeds a preset satisfaction level. For each port tallying operation scenario, the system uses the collected historical port tallying operation data to perform dependency analysis on the data of each operation node, determine the association between the operation nodes, and construct a port tallying operation directed graph. Based on the constructed port tallying operation directed graph, a node completion time window prediction model is trained and generated. According to the matching monitoring strategy, the real-time operation data of each operation node in the current operation scenario is obtained, and the corresponding trained node completion time window prediction model is input into it to obtain the predicted completion time of each operation node. The predicted completion time of each job node in the current job scenario is compared with the preset completion time of the corresponding job node, and a job node progress warning is issued for job nodes that exceed the preset completion time of the corresponding job node.

2. The port tally operation management method according to claim 1, characterized in that: The division of port tallying operation scenarios and the definition and quantification of node characteristics of each operation node in each operation scenario include: Classify port tallying operation scenarios based on port type, operation process type, and cargo type, including standard container loading and unloading scenarios at international or domestic ports, bulk cargo loading and unloading scenarios at international or domestic ports, and hazardous materials loading and unloading scenarios at international or domestic ports; define core operation nodes for different operation scenarios, including counting, inspection, classification, marking, damage inspection, and preparation of tally documents; define and quantify the core node characteristics of each operation node in each operation scenario, including: the number of operation node tasks, the complexity of operation node tasks, and the execution risk of operation node tasks, and obtain the quantitative standard values ​​of each operation node characteristic; The method of inputting the job node characteristics determined for each job node in the current job scenario into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current job node also includes: introducing a dynamic weight mechanism into the dynamic monitoring strategy model, setting core node feature weight parameters that match different core job nodes, adding feature weight parameters to each input job node feature, and then obtaining a monitoring strategy that matches the current job node.

3. The port tallying operation management method according to claim 1, characterized in that: Also includes: For each job node that exceeds the preset completion time of the corresponding job node, calculate the time deviation rate; Based on the calculated time deviation rate, the preset time deviation rate range interval of the corresponding operation scenario is determined, and the warning level is matched according to the determined time deviation rate range interval; wherein, each port tallying operation scenario is set with a preset time deviation rate range interval of the corresponding operation scenario, and each set time deviation rate range interval is preset with a warning level matching it; Emergency strategies are preset according to the warning level combination of operation scenarios and operation nodes, including: operation resource allocation strategy, operation process optimization strategy and operation auxiliary technology intervention strategy; for the current port tallying operation scenario, the preset emergency strategies are matched and executed for operation nodes at different warning levels.

4. The port tallying operation management method according to claim 3, characterized in that: Also includes: For each port tallying operation scenario, define the emergency events during the port tallying operation and the operational risk data caused by the emergency events, including: risk type, transmission probability and time delay coefficient; Risk transmission edge modeling includes defining edge attributes such as risk type, transmission probability, and time delay coefficient. Risk transmission edges are superimposed on the original directed graph dependency edges to form a dynamic extended graph, enabling updates to the original port tallying operation directed graph. Based on the updated directed graph of port tallying operations, a risk-corrected node completion time window prediction model is trained and generated to obtain the risk-corrected node completion time by utilizing the collected historical port tallying operation data, historical operating environment data, historical operating risk data obtained based on the historically collected port tallying operation data and operating environment data, and the historical risk-corrected node completion time corresponding to each operation node.

5. The port tally operation management method according to claim 4, characterized in that: Also includes: Preset emergency strategies that match the combination of operation scenario, risk type and operation node warning level to replace the preset emergency strategies that match the combination of operation scenario and operation node warning level; and for each combination of operation scenario, risk type and operation node warning level, a collective emergency strategy is preset, and the elements of the preset collective emergency strategy are emergency strategies with different response time efficiencies and are respectively set with a matching comprehensive risk index range interval, and the emergency strategy with a lower response time efficiencies has a smaller value of the comprehensive risk index range interval; Using deep learning algorithms, we can determine the risk impact and risk response difficulty of each operation node based on the operation data and operation environment data obtained from the early warning, and calculate the comprehensive risk index. The comprehensive risk index is calculated based on the weighted risk impact and risk response difficulty; In the current port tallying operation scenario, the preset collective emergency strategy is matched with the operation node based on the obtained risk type and warning level combination; the emergency strategy is continued to be matched according to the comprehensive risk index range interval of the calculated risk comprehensive index.

6. The port tallying operation management method according to claim 1, characterized in that: Also includes: Provides a job node progress visualization window to display the job progress of each job node in real time; Provides an interface that supports users in adjusting the task execution process of job nodes. It receives the task execution process of the job nodes adjusted by users and verifies whether it is inconsistent with the mandatory task execution process of the job nodes. If the verification is inconsistent, it will generate a prompt indicating that the task execution process adjustment of the job node has failed. Alternatively, the task execution process of each job node is automatically adjusted based on the progress percentage or expected completion time of each job node, so as to give priority to the task execution process corresponding to the job node with a small progress percentage or a later expected completion time, and the adjusted job node execution process does not conflict with the mandatory task execution process of the job node.

7. The port tallying operation management method according to claim 1, characterized in that: The monitoring method includes a centralized monitoring method or a distributed monitoring method; the monitoring resource equipment includes: Internet of Things sensors, 5G real-time transmission equipment, AR auxiliary equipment, AI visual equipment, scanning devices; the monitoring frequency includes: real-time monitoring, periodic monitoring or triggered monitoring.

8. A port tally operation management system, characterized in that: include: The port tallying operation definition module is used to divide the port tallying operation scenarios, define and quantify the node characteristics of each operation node in each operation scenario; The port tallying operation monitoring strategy acquisition module is used to build a dynamic monitoring strategy model. The operation node characteristics determined for each operation node in the current operation scenario are input into the dynamic monitoring strategy model to obtain a monitoring strategy that matches the current operation node, including monitoring resources and equipment, monitoring methods, and monitoring frequency. The dynamic monitoring strategy model uses a deep learning algorithm to train and generate historical monitoring strategies based on the node characteristics determined for each operation node in each operation scenario in historical port tallying operation data, as well as historical monitoring strategies where user satisfaction with the monitoring strategy adopted for the corresponding operation node exceeds a preset satisfaction level. The port tallying operation completion time prediction module is used to analyze the dependency relationships of the data of each operation node based on the collected historical port tallying operation data for each port tallying operation scenario, determine the association relationship between the operation nodes, construct a port tallying operation directed graph, and train and generate a node completion time window prediction model based on the constructed port tallying operation directed graph. According to the matching monitoring strategy, the real-time operation data of each operation node in the current operation scenario is obtained, and the corresponding trained node completion time window prediction model is input into it to obtain the predicted completion time of each operation node. The port tallying operation progress warning module is used to predict the completion time of each operation node in the current operation scenario, compare it with the preset completion time of the corresponding operation node, and issue an operation node progress warning for the operation node that exceeds the preset completion time of the corresponding operation node.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer device, characterized in that: The computer device includes a memory, a processor, and a program stored and executable on the memory, and when the program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images