A wharf sea side operation deduction method and system based on a multi-agent system

Abstract

This application provides a method and system for simulating port operations on the seaside based on a multi-agent system. Applied to the field of port terminal operation and management technology, the system constructs a multi-agent collaborative simulation system based on actual terminal resource allocation. Terminal physical equipment is abstracted into agents encapsulated with attributes, states, and behavioral logic. The task manager agent parses ship operation tasks and assigns tasks according to regional matching rules. Quay cranes, yard cranes, and container trucks receive tasks through asynchronous communication, make autonomous decisions based on their own resource states, and determine the task execution sequence using a dynamic action time prediction model to advance the simulation's logical timeline. The container trucks respond to commands, simulating container flow between quay cranes and the yard. During the simulation, key commands and status data are written to the database in real time, forming a traceable data trajectory. This supports real-time monitoring and post-event multi-dimensional indicator analysis, achieving distributed, dynamic, and data-driven high-fidelity simulation of the seaside operation process.

Description

Technical Field

[0001] This application relates to the field of port terminal operation and management technology, specifically to a terminal seaside operation simulation method and system based on a multi-agent system. Background Technology

[0002] The seaside operations of container terminals are a core element determining terminal efficiency, involving the coordinated operation of various equipment such as quay cranes, trucks, yard cranes, and forklifts. Currently, most mainstream terminals use centralized dispatch systems for coordination and command among equipment. However, this model has gradually revealed many problems in actual operation: lagging information exchange, poor response flexibility, and frequent resource conflicts. It is difficult to adapt to the real-time adjustment needs of dynamic operating environments. Especially during peak ship loading and unloading periods, the task connection between quay cranes and yard cranes often results in waiting, congestion, or empty runs for trucks due to scheduling mismatches. The mismatch in forklift operation rhythms further exacerbates the operational bottlenecks and seriously restricts the improvement of overall efficiency.

[0003] To overcome these bottlenecks, simulation systems and scheduling methods for various port operations have emerged. However, most existing solutions suffer from problems such as static models, centralized decision-making, and a disconnect between simulation and control. For example, some simulation systems can only perform offline simulations, with the simulation model and the actual production control system operating independently. They cannot synchronize the status of on-site equipment in real time, nor can they perform online simulations and dynamic responses to sudden disturbances such as equipment failures and task changes, lacking closed-loop linkage capabilities with the actual operation system. Some methods, while introducing multi-agent architectures, focus on static resource configuration optimization, with agents primarily evaluating strategies, lacking the ability to perform detailed simulations and forward-looking predictions of dynamic operation processes. Some methods attempt to integrate agent simulation with discrete event modeling, achieving basic process simulation capabilities, but the simulation process still relies on the assumption of fixed loading and unloading action durations, resulting in insufficient time prediction accuracy. Furthermore, the model's characterization of container positions is limited to the regional level, easily leading to misjudgments of yard conflicts and deviations in task coordination, ultimately resulting in significant errors between the simulation results and actual operations. These limitations make it difficult for existing solutions to meet the demands of modern ports for real-time performance, dynamic response, and multi-equipment collaborative simulations.

[0004] Based on this, in the face of the increasingly complex terminal operation environment and the management requirements of lean operation, a new terminal seaside operation simulation scheme is needed. Summary of the Invention

[0005] In view of this, the embodiments of this specification provide a terminal seaside operation simulation method and system based on a multi-agent system. By constructing a distributed collaborative mechanism for intelligent agents such as quay cranes, container trucks, and yard cranes, and combining dynamic action time prediction based on working conditions and refined spatiotemporal modeling, dynamic and forward-looking operation simulation of the container flow process between ships and yards can be realized.

[0006] The embodiments in this specification provide the following technical solutions:

[0007] This specification provides an embodiment of a method for simulating wharf seaside operations based on a multi-agent system, including:

[0008] A distributed multi-agent system is constructed, comprising: a task manager agent, a quay crane agent, a yard crane agent, and a truck agent; wherein the quay crane agent, yard crane agent, and truck agent correspond to the physical equipment at the terminal and are encapsulated with corresponding equipment attributes, resource status, and behavioral logic.

[0009] The task manager intelligent agent obtains a list of ship operation tasks, parses the task information, and based on the task information, simulates and assigns the tasks to the corresponding quay crane intelligent agent or yard crane intelligent agent in the form of logical task packages according to the preset operation area matching rules.

[0010] The quay crane agent, yard crane agent, and truck agent receive tasks through an asynchronous communication mechanism and, based on their own resource status, autonomously make decisions and simulate the execution of the assigned tasks.

[0011] When each intelligent agent simulates the execution of a task, the dynamic action time prediction model based on the working conditions determines the logical timing of the task execution in order to advance the logical timeline of the system.

[0012] The truck intelligent agent, in response to the interactive commands issued by the quay crane intelligent agent or the yard crane intelligent agent during the task simulation execution, simulates the container transfer process between the quay crane and the yard.

[0013] Key data returned based on preset key instructions during the simulation process are stored in the database to form a traceable operation data trajectory.

[0014] This specification also provides an embodiment of a terminal seaside operation simulation system based on a multi-agent system, including:

[0015] The system construction module is used to build a distributed multi-agent system, which includes: a task manager agent, a quay crane agent, a yard crane agent, and a truck agent; wherein the quay crane agent, yard crane agent, and truck agent correspond to the physical equipment of the terminal and encapsulate the corresponding equipment attributes, resource status, and behavioral logic.

[0016] The task management module is used to obtain a list of ship operation tasks through the task manager intelligent agent, parse the task information, and based on the task information, the task manager intelligent agent simulates and assigns the tasks to the corresponding quay crane intelligent agent or yard crane intelligent agent in the form of logical task packages according to the preset operation area matching rules.

[0017] The simulation module enables the quay crane agent, yard crane agent, and truck agent to receive tasks through an asynchronous communication mechanism, and to autonomously make decisions and simulate the execution of the assigned tasks based on their own resource status.

[0018] When each intelligent agent simulates the execution of a task, a dynamic action time prediction model based on the working conditions is used to determine the logical timing of the task execution, so as to advance the logical time axis of the system deduction.

[0019] The truck intelligent agent, in response to the interactive commands issued by the quay crane intelligent agent or the yard crane intelligent agent during the task simulation execution, simulates the container transfer process between the quay crane and the yard.

[0020] The data storage module is used to store key data returned based on preset key instructions during the simulation into the database to form a traceable operation data trajectory.

[0021] Compared with the prior art, the beneficial effects that at least one technical solution adopted in the embodiments of this specification can achieve include at least:

[0022] This application constructs a multi-agent collaborative simulation system based on actual terminal resource allocation, enabling high-precision, forward-looking dynamic simulation of seaside operations within a specific future time period (e.g., 4 hours). Each agent autonomously decides to execute tasks based on the attributes of packaging equipment and resource status. Combined with a dynamic action time prediction model, the simulation sequence is advanced, accurately simulating the flow of containers between quay cranes and the yard. Key data is recorded in real time to form a traceable operation trajectory. This allows for effective prediction of plan feasibility before actual operations begin, optimization of equipment coordination rhythm, and significantly improved foresight, robustness, and execution reliability of terminal operation plans. It provides high-fidelity, quantifiable, and optimizable pre-simulation decision support for production scheduling. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of a simulation method for wharf seaside operations in this application;

[0025] Figure 2 This is a flowchart of the intelligent agent interaction process for the unloading process in this application;

[0026] Figure 3 This is a flowchart of the intelligent agent interaction process for the loading process in this application;

[0027] Figure 4 This is a schematic diagram of the command prediction algorithm using a field bridge as an example in this application;

[0028] Figure 5 This is a framework diagram of a single intelligent agent in this application;

[0029] Figure 6 This is a diagram of a simulation system architecture for wharf seaside operations in this application. Detailed Implementation

[0030] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0031] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0032] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0033] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0034] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.

[0035] With the increasing intelligence and automation of ports, simulation and deduction technologies are being used more and more widely in optimizing terminal operations. However, most mainstream terminals currently use centralized scheduling systems for coordination and command between equipment. All information must be gathered at the center, and all instructions are issued uniformly from the center. When faced with situations such as dense ship arrivals, temporary adjustments to work plans, or sudden equipment failures, the system struggles to achieve rapid response and dynamic adjustment. It often causes a series of chain reactions, such as equipment waiting and work line congestion, due to delayed instructions, which seriously restricts further improvement in terminal operation efficiency. In addition, most existing simulation systems are still in offline deduction mode. Although they can replay historical operations or perform idealized simulations based on fixed parameters (such as fixed quay crane loading and unloading times), they ignore the fact that the duration of each operation in actual operations is dynamically affected by various factors such as container weight, specific location, and real-time wind speed. This results in existing offline simulation results deviating significantly from the actual situation on site, and cannot effectively support production decisions.

[0036] In view of this, the inventors, through research and improvement exploration, discovered that one of the key reasons for the distortion of existing terminal seaside operation simulation methods is insufficient time prediction accuracy. Existing methods use fixed time assumptions for each action, without considering factors such as the actual operating status of equipment, path differences, and dynamic interference. For example, the simulation model assumes that each action of the quay crane takes 3 minutes, while in actual operation it may fluctuate between 2 and 4 minutes. This causes the simulation results to accumulate deviations from the start of the task, making the progress of the entire simulation timeline gradually deviate from reality, and ultimately rendering the simulation results meaningless in practice. Another important reason is the coarse-grained problem of spatial modeling. There is a lack of modeling ability for fine-grained occupation of spatial resources. Existing simulation schemes can only locate containers in broad areas such as "Area A", and cannot be accurate to specific "bay position" or "layer position" level coordinates. This makes it impossible to accurately estimate the actual movement distance of the yard crane, and even more impossible to effectively predict the resource competition of different tasks for the same lane or storage location. Ultimately, this makes it difficult for existing simulation methods to achieve dynamic prediction, conflict early warning, and adaptive adjustment of seaside operation processes.

[0037] Based on this, the embodiments of this specification propose a terminal seaside operation simulation method based on a multi-agent system. The overall idea is as follows: by constructing a multi-agent collaborative simulation system based on actual terminal resource configuration, terminal physical equipment is abstracted into agents encapsulated with attributes, states, and behavioral logic. The task manager agent is responsible for parsing ship operation tasks and assigning tasks according to regional matching rules. Each device agent receives tasks through an asynchronous communication mechanism, makes autonomous decisions to execute based on its own resource status, and determines the task execution sequence according to a dynamic action time prediction model to advance the system simulation. During this process, the truck agent responds to interactive commands to simulate the container flow between the quay crane and the yard. At the same time, the system stores key simulation data in the database to form a traceable operation data trajectory, thereby realizing distributed, dynamic, and data-driven collaborative simulation of the terminal seaside operation process. This solves technical problems such as low efficiency of multi-equipment collaboration, mismatch of operation rhythm, and weak response capability to abnormal disturbances, and significantly improves the intelligence, adaptability, and overall operating efficiency of terminal operations.

[0038] The technical solutions provided by the various embodiments of this application are described below with reference to the accompanying drawings.

[0039] like Figure 1 As shown in the embodiments of this specification, a method for extrapolating wharf seaside operations based on a multi-agent system is provided, including:

[0040] Step S100: Construct a distributed multi-agent system, which includes: a task manager agent, a quay crane agent, a yard crane agent, and a truck agent; wherein the quay crane agent, the yard crane agent, and the truck agent correspond to the physical equipment of the terminal and encapsulate corresponding equipment attributes, resource status, and behavioral logic.

[0041] Specifically, the key operational equipment on the sea side of the wharf is abstracted into multiple independently operating intelligent agents. During the system initialization phase, several quay crane intelligent agents, yard crane intelligent agents, and container truck intelligent agents are created. Each intelligent agent loads configuration parameters such as its operating area, initial position, and lane resource status.

[0042] Step S200: The task manager agent obtains the list of ship operation tasks, parses the task information, and based on the task information, the task manager agent simulates and assigns the tasks to the corresponding quay crane agent or yard crane agent in the form of logical task packages according to the preset operation area matching rules.

[0043] Specifically, the task manager agent obtains the list of ship operation tasks from the terminal operating system (TOS) and parses key task information such as container number, operation direction (loading or unloading), original tank position, target tank position, and ship berthing position from the list of operation tasks.

[0044] During task allocation, based on the parsing results, the task manager assigns tasks to the corresponding quay crane or yard crane agent in the form of logical task packages according to the regional matching rules, and finds idle container truck agents for transportation connection. Tasks are transferred between agents in the form of task packages, without including physical motion model calculations, thereby ensuring that the simulation process focuses on operational logic and resource coordination.

[0045] Step S300: The quay crane agent, the yard crane agent, and the truck agent receive tasks through an asynchronous communication mechanism, and autonomously make decisions and simulate the execution of the assigned tasks based on their own resource status.

[0046] When each intelligent agent simulates the execution of a task, the dynamic action time prediction model based on the working conditions determines the logical timing of the task execution in order to advance the logical timeline of the system.

[0047] The truck intelligent agent, in response to the interactive commands issued by the quay crane intelligent agent or the yard crane intelligent agent during the task simulation execution, simulates the flow of containers between the quay crane and the yard.

[0048] Specifically, the task manager agent acts as the system coordination hub, driving each agent to receive tasks, update status, and respond collaboratively through an asynchronous communication mechanism, forming a loosely coupled deduction system.

[0049] In this system, the quay crane agent and the yard crane agent simulate container retrieval and ship loading and unloading operations, respectively, and manage the occupation and release of their respective operating lanes; the truck agent is responsible for simulating the horizontal transportation of containers between the quay crane and the yard crane, and supports task queuing and state transition; the task manager agent organizes the task process according to the operation direction (loading or unloading), and drives each agent to execute the simulation according to the actual operation logic sequence.

[0050] Step S400: Store the key data returned based on the preset key instructions during the simulation into the database to form a traceable operation data trajectory.

[0051] Specifically, during the simulation operation, the built-in data writing module writes the instructions and status changes of each stage into the background database in real time. The written content covers key information such as task allocation time, equipment action type (such as lifting, moving, handover), task execution start and end time, equipment location information, resource occupation and release records, and task completion status, thereby constructing a complete and traceable operation data trajectory, providing a reliable data foundation for subsequent multi-dimensional analysis and evaluation.

[0052] This application uses a multi-agent parallel inference mechanism to simulate the autonomous response and collaborative behavior of equipment in a dynamic environment. Combined with the ability to record data throughout the process and analyze indicators in two stages, it constructs a closed-loop technical system of "inference-recording-evaluation".

[0053] In some embodiments, the task information includes at least: the container's operation type, original location, target location, and ship berth coordinates; wherein the original location and target location are accurate to the ship berth or yard berth level;

[0054] If the operation type is unloading, then it is deduced that the quay crane agent performs the unloading action from the original position, the container truck agent transports the container from the quay crane area to the yard area where the target location is located, and the yard crane agent places the container at the target location.

[0055] If the operation type is loading onto a ship, it is deduced that the yard crane agent performs a container retrieval action from the original position, the container truck agent transports the container from the yard crane area to the quay crane area corresponding to the ship berth coordinates, and the quay crane agent loads the container to the target location.

[0056] During implementation, the simulation process takes the list of ship operation tasks as input. The task manager agent parses the operation type, original location, target location and ship berth coordinates of each container to generate an initial task queue.

[0057] For unloading tasks, according to Figure 2 The process shown illustrates the following key steps in sequence: The quay crane agent performs the unloading action from the ship from its original position and sends instructions to the container truck agent. The container truck agent, based on the received instructions, horizontally transports the container from the quay crane operation area to the yard area where the target location is located, and sends instructions to the yard crane agent. The yard crane agent accurately places the container at the target location, completes the unloading process, and notifies the task management agent.

[0058] For loading tasks, according to Figure 3 The process is illustrated as follows: The yard crane agent retrieves the container from the yard based on its original location and sends a command to the truck agent. The truck agent then transports the container to the quay crane operation area corresponding to the ship's berth coordinates and sends a command to the quay crane agent. Upon receiving the command, the quay crane agent loads the container to the target location on the ship, completing the loading process, and notifies the task management agent.

[0059] It should be noted that all locations in the system are described with the precision of ship container positions (Bay / Row / Tier) or yard container positions (stack area / bay / layer), which realizes fine-grained modeling of space resource occupation, effectively avoiding task conflicts and connection deviations caused by positioning ambiguity. By combining the operation type and precise spatial coordinates, the system drives each agent to execute standardized operation sequences, ensuring the accuracy and reliability of the simulation process.

[0060] In some embodiments, the dynamic action time prediction model obtains the expected action execution time and state change instructions by calling an external simulation service interface;

[0061] The estimated execution time of the action is generated based on the current state of the device, the scheduling strategy, and the current operating conditions.

[0062] During implementation, such as Figure 4 As shown, a dynamic action time prediction model is introduced to establish an instruction response mechanism consistent with the actual production control logic. When a quay crane agent receives a box retrieval task, it sends an action request to the external simulation service interface. This interface does not return a fixed empirical value, but calculates and returns a dynamic action execution time (e.g., 15 seconds) and corresponding state change instruction in real time based on the current state of the equipment, scheduling strategy, and specific operating parameters at the time of the request. The simulation system advances the logical timeline based on this precise time value and updates the equipment state synchronously, thereby ensuring the accuracy of time prediction at the basic unit level.

[0063] All task instructions and status feedback generated throughout the simulation process are modeled strictly according to the control timing and communication logic of the real production system. This ensures that the interaction flow between intelligent agents is highly consistent with the instruction transmission and response timing in actual operations, guaranteeing that the simulation process itself is a reliable mapping of the real system's operating logic. Based on the aforementioned high-precision time prediction and realistic process simulation, each intelligent agent (quay crane, yard crane, and container truck) can autonomously simulate highly realistic operational behaviors and resource competition according to real constraints such as the number of equipment, lane capacity, movement speed, and task priority, combined with a dynamic action time prediction model based on operating conditions. This allows for accurate prediction of continuous task execution timing, changes in equipment occupancy status, and proactive identification of potential congestion points and operational bottlenecks.

[0064] In some embodiments, the key instructions interact with external systems through an encrypted interface to return key data;

[0065] The data transmission with external systems uses the HTTPS protocol and sensitive fields are encrypted; a security authentication mechanism is enabled when the key data is stored in the database.

[0066] During implementation, such as Figure 5 and Figure 6 As shown, all key instructions during task execution, including task allocation, action initiation, resource usage, and task completion, interact with external systems through encrypted API interfaces, and the returned execution time, path instructions, device response, and other information are written to the background database in real time.

[0067] Data transmission uses the HTTPS protocol and sensitive fields are encrypted. Database connections are secured with a security authentication mechanism that strictly authenticates access identities. Only authorized services can read and write data, ensuring the integrity and confidentiality of the simulation data.

[0068] In some embodiments, the autonomous decision-making and simulated execution of the assigned task includes:

[0069] When the quay crane intelligent agent, the yard crane intelligent agent, or the truck intelligent agent receives a task request, it determines whether the current resources are occupied.

[0070] If resources are occupied, the system enters a queuing state until the resources are released, at which point the assigned task is simulated to simulate resource competition and task queuing behavior in real-world operations.

[0071] Specifically, such as Figure 5 As shown, when an agent receives a task request assigned by the task manager, the agent's identity and status module provides the latest encapsulated resource status. The behavior logic module makes real-time judgments based on the latest encapsulated resource status to detect whether the key resources on which the agent depends for executing the task have been occupied. For example, the quay crane agent judges whether the working lane is free; the truck agent judges whether the vehicle is in an idle state.

[0072] If the behavior logic module determines that the resource is occupied, it adds the task to its own task waiting queue and enters the queuing waiting state, and continuously monitors the resource status. Once the resource is released, it immediately retrieves the task from the queue in order and starts to simulate execution, which accurately simulates the task queuing phenomenon caused by the competition for limited physical resources such as equipment, lanes, and container spaces in real terminal operations.

[0073] It should be noted that the intelligent agent includes the following modules: identity and state module, behavior logic module, communication interface module, and data writing module;

[0074] The identity and state module is mainly responsible for the identification and state management of the intelligent agent; the behavior logic module mainly handles the behavior strategy and decision logic of the intelligent agent; the communication interface module mainly provides the interface for the intelligent agent to communicate with other systems or environments; and the data writing module is mainly responsible for writing data into the corresponding storage or processing system.

[0075] In some embodiments, state data of each agent is collected in real time during the simulation process;

[0076] Based on the status data, real-time operating indicators are generated through the real-time indicator analysis module; wherein, the operating indicators include at least one of the following: current equipment load rate, task completion rate, truck utilization rate, and lane occupancy rate;

[0077] Based on the real-time operating metrics, potential bottleneck agents are identified, and corresponding conflict events are recorded.

[0078] Specifically, during the simulation process, the system continuously collects status data of each intelligent agent (such as quay cranes, yard cranes, and container trucks), and dynamically calculates key operating indicators, such as equipment load rate, task completion rate, and container truck turnover frequency, through the real-time indicator analysis module.

[0079] For example, when the system detects that the task queue of a certain bridge agent is continuously growing and the load rate exceeds the preset threshold, the system automatically identifies the agent's equipment as a potential operational bottleneck, records conflict events, and writes the key data generated during the simulation process into the database in real time, thereby supporting the full-process tracking of each device and each task within the simulation cycle.

[0080] In some embodiments, the method for simulating wharf seaside operations further includes:

[0081] During the simulation, when the task manager agent detects that the quay crane agent, the yard crane agent, or the truck agent has not returned a status response within a preset time, it determines that the task has failed and initiates a task reassignment request.

[0082] The strategy for the reallocation request includes: removing currently unresponsive agents, reallocating the current task according to the scheduling strategy, and continuing the deduction based on the new allocation result.

[0083] Specifically, during the execution of a task, a truck intelligent agent simulates a communication interruption. If the task manager intelligent agent detects that the device has not returned a status response within the expected time, it determines that the task has failed and initiates a task reassignment request to an external system. Another truck intelligent agent takes over the original transportation task. After receiving the updated transportation instructions, the new truck intelligent agent obtains the new execution time and path parameters through the interface. The system continues to deduce the subsequent operation process based on the new parameters, simulating the scenario where a backup device takes over after a device failure in actual operation. This achieves seamless task connection and rapid process recovery, significantly enhancing the simulation system's autonomous response and adaptability to dynamic disturbance events.

[0084] In some embodiments, the method for simulating wharf seaside operations further includes:

[0085] After the simulation is completed, an overall evaluation index is generated based on the operation data trajectory to quantitatively evaluate the execution efficiency and resource balance of the operation plan.

[0086] The overall evaluation indicators include at least one of the following: total operation time, average task turnaround time, resource waiting time, and equipment utilization rate.

[0087] Specifically, after the simulation, the overall indicator analysis module performs a global summary and in-depth analysis of the simulation cycle based on the full-process instruction records stored in the database, generating a series of comprehensive evaluation indicators, such as average task duration, resource waiting time, and equipment utilization rate. From multiple dimensions such as efficiency, balance, and reliability, the module accurately quantifies and compares the execution effects of different work plan schemes, thereby optimizing the work plan.

[0088] In some embodiments, during the simulation process, the simulation data generated is continuously pushed to the visualization platform through a secure message queue;

[0089] The secure message queue enables transmission encryption and identity authentication mechanisms for communication.

[0090] During implementation, all simulation results are pushed to the visualization platform through a secure transmission message queue. The queue communication uses transmission encryption and identity authentication to ensure that the information is not illegally obtained.

[0091] Specifically, such as Figure 6 As shown, a dedicated, asynchronous data channel between the system and the visualization platform is established through a secure transmission message queue. As the simulation logic timeline progresses, the system encapsulates various simulation data, such as real-time device location, task status changes, resource usage, real-time operating indicators, and early warning events, into messages and continuously pushes them to the message queue in real time. To prevent data from being stolen or tampered with during transmission, the message queue uses a transmission encryption mechanism throughout the communication process and performs strict two-way authentication on the message producer (simulation system) and consumer (visualization platform). This effectively eliminates the risks of information leakage, tampering, and unauthorized access while enabling real-time visual monitoring of the operation status.

[0092] Compared to existing technologies, this application achieves breakthroughs in simulation accuracy, dynamic adaptability, and system practicality: by introducing a dynamic time prediction model based on operating conditions, it significantly improves the accuracy of predicting the action time of equipment such as quay cranes, container trucks, and yard cranes; by pinpointing container positions to the ship and yard levels, it achieves fine-grained modeling of space resource occupancy, effectively avoiding misjudgments of yard conflicts; the simulation instructions strictly follow the timing and communication logic of the real control system, ensuring that the instruction flow conforms to the actual operating process; it supports the dynamic injection of disturbance events and autonomous rescheduling of intelligent agents, enhancing the response capability to abnormal operating conditions. Compared with existing simulation systems, it not only achieves refined and distributed logical simulation of the offshore operation process, but also, through high-fidelity modeling and database-supported multi-dimensional indicator analysis, makes the simulation results more realistic, measurable, comparable, and optimizable, effectively improving the scientific nature, adaptability, and executability of terminal operation plans.

[0093] This application supports the dynamic injection of disturbance events (such as equipment failure and task changes), and each intelligent agent can respond and reschedule autonomously based on local information, thereby improving its adaptability to abnormal operating conditions.

[0094] Based on the same inventive concept, this application also provides a terminal seaside operation simulation system based on a multi-agent system, comprising:

[0095] The system construction module is used to build a distributed multi-agent system, which includes: a task manager agent, a quay crane agent, a yard crane agent, and a truck agent; wherein the quay crane agent, yard crane agent, and truck agent correspond to the physical equipment of the terminal and encapsulate the corresponding equipment attributes, resource status, and behavioral logic.

[0096] The task management module is used to obtain a list of ship operation tasks through the task manager intelligent agent, parse the task information, and based on the task information, the task manager intelligent agent simulates and assigns the tasks to the corresponding quay crane intelligent agent or yard crane intelligent agent in the form of logical task packages according to the preset operation area matching rules.

[0097] The simulation module enables the quay crane agent, yard crane agent, and truck agent to receive tasks through an asynchronous communication mechanism, and to autonomously make decisions and simulate the execution of the assigned tasks based on their own resource status.

[0098] When each intelligent agent simulates the execution of a task, a dynamic action time prediction model based on the working conditions is used to determine the logical timing of the task execution, so as to advance the logical time axis of the system deduction.

[0099] The truck intelligent agent, in response to the interactive commands issued by the quay crane intelligent agent or the yard crane intelligent agent during the task simulation execution, simulates the container transfer process between the quay crane and the yard.

[0100] The data storage module is used to store key data returned based on preset key instructions during the simulation into the database to form a traceable operation data trajectory.

[0101] This application is based on the simulation process data recorded in the database. The system integrates an indicator analysis module, which supports two types of analysis functions: First, real-time indicator analysis, which dynamically calculates the current equipment load rate, task completion progress, truck utilization rate, lane occupancy rate and other operating status indicators during the simulation process, helping users to identify potential bottlenecks in real time; Second, post-operation overall indicator analysis, which summarizes the full-cycle data after the simulation ends and generates comprehensive evaluation indicators such as total operation time, average task turnaround time, equipment idle time, number of task waiting times, and frequency of resource conflicts, which are used to quantify the execution efficiency and resource balance of different operation plan schemes.

[0102] In this specification, the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the descriptions of the embodiments described later are relatively simple, and relevant parts can be referred to the descriptions of the foregoing embodiments.

[0103] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for extrapolating wharf seaside operations based on a multi-agent system, characterized in that, include: A distributed multi-agent system is constructed, comprising: a task manager agent, a quay crane agent, a yard crane agent, and a truck agent; wherein the quay crane agent, yard crane agent, and truck agent correspond to the physical equipment at the terminal and are encapsulated with corresponding equipment attributes, resource status, and behavioral logic. The task manager intelligent agent obtains the list of ship operation tasks from the terminal operating system, parses the task information, and based on the task information, simulates and assigns the tasks to the corresponding quay crane intelligent agent or yard crane intelligent agent in the form of logical task packages according to the preset operation area matching rules. The quay crane agent, yard crane agent, and truck agent receive tasks through an asynchronous communication mechanism and, based on their own resource status, autonomously make decisions and simulate the execution of the assigned tasks. When each intelligent agent simulates the execution of a task, a dynamic action time prediction model based on the working conditions determines the logical timing of the task execution to advance the logical timeline of the system. The dynamic action time prediction model is used to dynamically generate the expected execution time of the specific task based on the real-time working conditions at the time of the specific task execution before each intelligent agent executes the specific task, so as to establish an instruction response mechanism consistent with the actual production control logic. The truck intelligent agent, in response to the interactive commands issued by the quay crane intelligent agent or the yard crane intelligent agent during the task simulation execution, simulates the container transfer process between the quay crane and the yard. Key data returned based on preset key instructions during the simulation process are stored in the database to form a traceable operation data trajectory.

2. The wharf seaside operation simulation method according to claim 1, characterized in that, The task information includes at least: the container operation type, original location, target location, and ship berth coordinates; wherein the original location and target location are accurate to the ship berth or yard berth level; If the operation type is unloading, then it is deduced that the quay crane intelligent body performs the unloading action from the original position, the container truck intelligent body transports the container from the quay crane area to the yard area where the target location is located, and the yard crane intelligent body places the container at the target location. If the operation type is loading onto a ship, it is deduced that the yard crane agent performs a container retrieval action from the original position, the container truck agent transports the container from the yard crane area to the quay crane area corresponding to the ship berth coordinates, and the quay crane agent loads the container to the target location.

3. The method for simulating wharf operations on the sea side according to claim 1, characterized in that, The dynamic action time prediction model obtains the expected action execution time and state change instructions by calling an external simulation service interface. The estimated execution time of the action is generated based on the current state of the device, the scheduling strategy, and the current operating conditions.

4. The method for simulating wharf seaside operations according to claim 1, characterized in that, The key instructions interact with external systems through an encrypted interface, returning key data; The data transmission with external systems uses the HTTPS protocol and sensitive fields are encrypted; a security authentication mechanism is enabled when the key data is stored in the database.

5. The wharf seaside operation simulation method according to claim 1, characterized in that, The autonomous decision-making and simulated execution of assigned tasks include: When the quay crane intelligent agent, the yard crane intelligent agent, or the truck intelligent agent receives a task request, it determines whether the current resources are occupied. If resources are occupied, the system enters a queuing state until the resources are released, at which point the assigned task is simulated to simulate resource competition and task queuing behavior in real-world operations.

6. The method for simulating wharf seaside operations according to claim 1, characterized in that, During the simulation, the state data of each intelligent agent is collected in real time; Based on the status data, real-time operating indicators are generated through the real-time indicator analysis module; wherein, the operating indicators include at least one of the following: current equipment load rate, task completion rate, truck utilization rate, and lane occupancy rate; Based on the real-time operating metrics, potential bottleneck agents are identified, and corresponding conflict events are recorded.

7. The method for simulating wharf seaside operations according to claim 1, characterized in that, The simulation method for wharf seaside operations also includes: During the simulation, when the task manager agent detects that the quay crane agent, the yard crane agent, or the truck agent has not returned a status response within a preset time, it determines that the task has failed and initiates a task reassignment request. The strategy for the reallocation request includes: removing currently unresponsive agents, reallocating the current task according to the scheduling strategy, and continuing the deduction based on the new allocation result.

8. The method for simulating wharf seaside operations according to claim 1, characterized in that, The simulation method for wharf seaside operations also includes: After the simulation is completed, an overall evaluation index is generated based on the operation data trajectory to quantitatively evaluate the execution efficiency and resource balance of the operation plan. The overall evaluation indicators include at least one of the following: total operation time, average task turnaround time, resource waiting time, and equipment utilization rate.

9. The method for simulating wharf seaside operations according to any one of claims 1-8, characterized in that, During the simulation, the simulation data generated is continuously pushed to the visualization platform through a secure message queue; The secure message queue enables transmission encryption and identity authentication mechanisms for communication.

10. A terminal seaside operation simulation system based on a multi-agent system, characterized in that, include: The system construction module is used to build a distributed multi-agent system, which includes: a task manager agent, a quay crane agent, a yard crane agent, and a truck agent; wherein the quay crane agent, yard crane agent, and truck agent correspond to the physical equipment of the terminal and encapsulate the corresponding equipment attributes, resource status, and behavioral logic. The task management module is used to obtain a list of ship operation tasks from the terminal operating system through the task manager intelligent agent, parse the task information, and based on the task information, the task manager intelligent agent simulates and assigns the tasks to the corresponding quay crane intelligent agent or yard crane intelligent agent in the form of logical task packages according to the preset operation area matching rules. The simulation module enables the quay crane agent, yard crane agent, and truck agent to receive tasks through an asynchronous communication mechanism, and to autonomously make decisions and simulate the execution of the assigned tasks based on their own resource status. When each intelligent agent simulates the execution of a task, a dynamic action time prediction model based on the working conditions determines the logical timing of the task execution, thereby advancing the logical timeline of the system. The dynamic action time prediction model is used to dynamically generate the estimated execution time of the specific task based on the real-time working conditions during the execution of the specific task before each intelligent agent executes the specific task, so as to establish an instruction response mechanism consistent with the actual production control logic. The truck intelligent agent, in response to the interactive commands issued by the quay crane intelligent agent or the yard crane intelligent agent during the task simulation execution, simulates the container transfer process between the quay crane and the yard. The data storage module is used to store key data returned based on preset key instructions during the simulation into the database to form a traceable operation data trajectory.

Images