Port container scheduling system and method based on GRASP and SAC

Abstract

The invention discloses a port container scheduling system and method based on GRASP and SAC. The system comprises a data acquisition module which is used for acquiring real-time data from port equipment and an environment sensor and preprocessing the real-time data; the optimization algorithm module is used for receiving the data output by the data acquisition module and generating a scheduling scheme, or receiving abnormal data fed back by the real-time monitoring module, automatically starting a GRASP algorithm and / or a SAC algorithm and regenerating the scheduling scheme; the task scheduling module is used for receiving the scheduling scheme and executing and adjusting specific task distribution and equipment scheduling operation; and the real-time monitoring module is used for monitoring the equipment state and the task execution condition and transmitting abnormal information to the optimization algorithm module. According to the method, the optimization algorithm modules are constructed by combining the GRASP and the SAC algorithm, and the modules are clear in division of labor and are in close cooperation, so that the whole scheduling system forms an efficient and cooperative whole, and the overall efficiency and stability of the scheduling system are improved.

Description

Technical Field

[0001] This invention relates to an intelligent management system for port logistics, and more particularly to a port container scheduling system and method based on GRASP and SAC collaborative optimization. Background Technology

[0002] In today's booming global trade, ports, as key hubs for cargo transportation, directly impact the overall efficiency of logistics through their container scheduling efficiency. Traditional port container scheduling methods, largely based on static rules and simple algorithms, reveal numerous problems when facing the complex and ever-changing port environment. For example, in unforeseen circumstances such as equipment failure, sudden changes in workload, or traffic congestion, traditional scheduling methods cannot adjust scheduling strategies in a timely and effective manner, leading to extended operation times, reduced equipment utilization, and increased energy consumption. Furthermore, traditional methods struggle to balance operational efficiency, energy consumption, and equipment utilization, often resulting in compromises. For instance, simply pursuing operational speed may lead to high-load operation of equipment, a surge in energy consumption, accelerated equipment wear, and a shortened lifespan. Simultaneously, multi-objective optimization has always been a challenge in port scheduling. Traditional methods struggle to find a balance between shortening loading and unloading times, reducing energy consumption, and improving equipment utilization, limiting the improvement of port operational efficiency. With the advancement of smart port construction, there is an urgent need for an innovative scheduling system and method that can adapt to the dynamic changes in ports and achieve multi-objective collaborative optimization. Therefore, a container scheduling system that achieves multi-objective optimization is crucial. Summary of the Invention

[0003] Purpose of the Invention: The purpose of this invention is to provide a port container scheduling system based on GRASP and SAC, addressing the problems of insufficient global optimization, poor dynamic adaptability, and unsatisfactory multi-objective optimization effects in existing port container scheduling technologies. Another objective of this invention is to propose an operation method for a port container scheduling system based on GRASP and SAC, addressing how to improve the efficiency, stability, and resource utilization of port container scheduling through this system.

[0004] Technical solution: The port container scheduling system based on GRASP and SAC described in this invention includes: The data acquisition module is used to collect real-time data from port equipment and environmental sensors and to preprocess it. The optimization algorithm module is used to receive data output from the data acquisition module and generate a scheduling scheme, or to receive abnormal data fed back by the real-time monitoring module, automatically start the global optimization process of the GRASP algorithm and / or the local optimization process of the SAC algorithm, and regenerate the scheduling scheme. The task scheduling module is used to receive scheduling plans and execute and adjust specific task allocation and equipment scheduling operations; The real-time monitoring module is used to monitor the device status and task execution, and transmit abnormal information to the optimization algorithm module.

[0005] Preferably, the data acquisition module includes: Position sensors are used to monitor the position and trajectory of equipment and containers; Load sensors are used to monitor equipment load and operating status; Environmental sensors are used to acquire port meteorological data; Traffic flow sensors are used for real-time data acquisition, predictive analysis, and closed-loop feedback. Equipment status sensors are used to monitor the operating status of critical equipment; A multi-channel data processing unit that supports multiple data interfaces is used to perform noise filtering, standardization, and anomaly detection on the collected data to ensure data quality and provide reliable input for the optimization algorithm module.

[0006] Preferably, in step (1), the optimization algorithm module includes a GRASP algorithm submodule for global optimization and a SAC algorithm submodule for local dynamic optimization; The GRASP submodule performs a global search in the multidimensional parameter space, and generates an initial scheduling scheme by constructing a greedy random solution and a local search, taking into account factors such as task priority, equipment allocation and loading / unloading order; The SAC submodule constructs a deep neural network optimization strategy based on the initial scheduling scheme generated by the GRASP submodule. Its state space includes equipment load, task progress, traffic flow and environmental state information, and its action space covers equipment task allocation and priority adjustment operations. It uses real-time data to dynamically adjust the local scheduling strategy and combines the Actor-Critic structure to optimize the scheduling scheme, thereby improving scheduling flexibility and adaptability.

[0007] Furthermore, the global optimization process of the GRASP algorithm includes: During the initialization phase, parameters are set, including the random factor used in the construction phase and the neighborhood search radius for local searches. During the construction phase, based on the port equipment's processing capacity and task time requirements, a combination of tasks and equipment is randomly selected to form an initial feasible solution; During the local search phase, the 2-opt method is used to optimize the initial solution, and the quality of the solution is evaluated by calculating the fitness function, which is defined as:

[0008] in, For the task The actual completion time, For the task Maximum allowed completion time For equipment Energy consumption per unit time For equipment Homework time, For equipment Total available time, For the task The delay time, For the task Maximum allowable delay time These are the weighting coefficients; When the maximum number of iterations is reached or the fitness value converges, the initial scheduling scheme is output.

[0009] Furthermore, the local optimization process of the SAC algorithm includes: Port scheduling tasks are modeled as Markov decision processes. In model construction and parameter definition, state space parameters are determined based on remaining equipment capacity, task urgency, and time windows. Action policies in the action space aim to maximize long-term cumulative rewards and policy entropy. In the training and updating of the deep reinforcement learning network, the Actor network outputs an entropy-regularized stochastic scheduling policy based on the current state. Critic network estimates state-action value function ; The reward function is defined as:

[0010] in, These are the weighting coefficients; At each time step, the system samples the current state, the Actor network outputs an action, and after execution, obtains the new state and immediate reward. The parameters of the Actor and Critic networks are updated using gradient descent. When the maximum number of iterations or the reward function convergence condition is met, the optimized local scheduling scheme is output.

[0011] In this invention, the GRASP algorithm performs optimization at the global level by constructing a greedy random solution and a local search process to find a high-quality initial scheduling scheme in a multi-dimensional parameter space. During the construction phase, factors such as the port equipment's processing capacity, task time requirements, and priorities are comprehensively considered, and a greedy random strategy is used to determine the initial allocation of tasks and equipment. In the local search phase, a specific optimization method, such as the 2-opt algorithm, is used to perform a neighborhood search on the initial solution, continuously optimizing the solution's quality to achieve comprehensive optimization of total operation time, equipment energy consumption, and task delay rate. The SAC algorithm, on the other hand, focuses on local optimization and dynamic adjustment. It models the port scheduling task as a Markov decision process, constructing a deep neural network to learn real-time port status information and dynamically optimize task allocation and equipment operation strategies. In this process, the SAC algorithm utilizes an Actor-Critic architecture. The Actor network outputs a stochastic scheduling strategy with entropy regularization, and the Critic network estimates the state-action value function. By maximizing cumulative rewards and policy entropy, it balances exploration and utilization, enabling the scheduling strategy to quickly adapt to changes in the port environment and improving scheduling flexibility and adaptability.

[0012] Preferably, the task scheduling module includes: The task parsing unit is used to receive the scheduling scheme from the optimization algorithm module and parse it into a sequence of device operation tasks; The instruction generation unit is used to generate a set of device control instructions based on the parsing results, including operation steps, execution order, time and device parameters; The communication control unit is used to transmit instructions to the equipment control system through various communication protocols, and to receive equipment execution feedback in real time. Based on the feedback, it adjusts task allocation to ensure efficient task execution.

[0013] The task scheduling module, based on the optimal scheduling scheme generated by the GRASP and SAC algorithms, translates the results of global and local optimizations into specific equipment instructions to execute actual container loading, unloading, and transportation operations. It decomposes the optimized loading and unloading sequence and equipment usage plan into instructions for quay cranes, yard cranes, and automated guided vehicles (AGVs), transmitting these instructions to the control systems of each device via various communication methods such as RS485, Ethernet, and 5G, ensuring that the equipment executes container loading and unloading tasks according to plan. The module continuously monitors the operational progress and task completion status of the equipment, and based on real-time feedback information, such as equipment status anomalies or task delays, it quickly adjusts the equipment scheduling scheme to ensure the efficient and continuous execution of container loading and unloading tasks, achieving optimal execution of scheduled tasks.

[0014] Preferably, the real-time monitoring module includes: The equipment status monitoring unit is used to monitor the operating status of equipment using sensors such as vibration, temperature, and current. The task progress tracking unit is used to record and analyze the task execution progress in real time. The anomaly handling unit is used to generate an anomaly report and feed it back to the optimization algorithm module when equipment failure, task delay or traffic congestion is detected, triggering the optimization process, adjusting the scheduling strategy, and ensuring stable system operation.

[0015] The real-time monitoring module comprehensively monitors equipment operation status, task execution progress, traffic flow changes, and environmental parameter fluctuations during port operations through real-time data acquisition and intelligent analysis technologies. Utilizing machine learning algorithms to build an anomaly detection model, it can quickly and accurately identify abnormal situations such as equipment failures, task delays, and traffic congestion. Once an anomaly is detected, a feedback mechanism is immediately activated, sending the anomaly information and related data to the intelligent decision-making optimization module, triggering a re-optimization process to ensure the real-time effectiveness of the scheduling plan and the continuity of operations.

[0016] Preferably, the system further includes a user interface module comprising: The data display unit is used to display the port equipment status, task execution status, scheduling plan and energy consumption data in real time. The control panel unit provides an operation interface, allowing users to adjust scheduling parameters and modify scheduling schemes. The emergency intervention unit is used to provide emergency operation options, such as emergency pause and equipment switching, in the event of system malfunction, thereby improving the system's emergency response capabilities.

[0017] The user interface module features an intuitive and user-friendly visual interface, providing port managers and operators with real-time operational information, including dynamic charts and map positioning displays of key indicators such as equipment status, task progress, and resource consumption. It supports online adjustment of scheduling strategy parameters, such as task priority weights and equipment resource allocation ratios, enabling flexible manual intervention. Simultaneously, it offers historical data query and in-depth analysis functions to help users summarize lessons learned, optimize long-term scheduling strategies, and improve the overall port management level. Based on the above technical solutions, real-time data acquisition and intelligent analysis technologies comprehensively monitor equipment operation status, task execution progress, traffic flow changes, and environmental parameter fluctuations during port operations. An anomaly detection model built using machine learning algorithms can quickly and accurately identify anomalies such as equipment failures, task delays, and traffic congestion. Once an anomaly is detected, a feedback mechanism is immediately activated, feeding the anomaly information and related data back to the intelligent decision optimization module, triggering a re-optimization process to ensure the real-time effectiveness of the scheduling plan and the continuity of operations.

[0018] The second aspect of this invention discloses a method for operating the aforementioned port container scheduling system based on GRASP and SAC, comprising the following steps: The data acquisition module collects port equipment, task, traffic and environmental data in real time, performs preprocessing such as noise reduction and standardization, and transmits the processed data to the optimization algorithm module through a standardized interface; The GRASP submodule in the optimization algorithm module analyzes the preprocessed data, searches in the multidimensional parameter space, and generates an initial scheduling scheme. The goal is to optimize the loading and unloading sequence, equipment allocation, and task priority, thereby reducing the total operation time, equipment energy consumption, and task delay rate. The SAC submodule in the optimization algorithm module is based on the initial scheduling scheme generated by the GRASP submodule. It combines real-time data to build a deep neural network, dynamically adjusts task allocation and loading / unloading order, responds to port emergencies and equipment status changes, and improves the flexibility and adaptability of the scheduling scheme. The task scheduling module transforms the optimized scheduling scheme into operation instructions and sends them to the equipment control system. The real-time monitoring module monitors task execution and equipment status to ensure that the scheduling scheme is implemented as planned. When the real-time monitoring module detects an anomaly, it feeds back the anomaly information to the optimization algorithm module, and adjusts the scheduling strategy in real time based on the feedback to ensure that container loading and unloading tasks are carried out continuously and efficiently. The user interface module displays scheduling information to operators, who can manually intervene or adjust scheduling parameters based on execution results and actual needs. The system further optimizes strategies based on feedback, providing a reference for subsequent scheduling tasks.

[0019] Preferably, when the real-time monitoring module detects an anomaly, it feeds back the anomaly information to the optimization algorithm module, and adjusts the scheduling strategy in real time based on the feedback, including: The real-time monitoring module continuously monitors equipment status, task progress, and traffic flow. When an anomaly occurs, the feedback mechanism transmits the anomaly information to the optimization algorithm module. Based on the feedback data, the optimization algorithm module automatically initiates the GRASP algorithm global optimization or the SAC algorithm local optimization process, recalculates the scheduling strategy, and adjusts equipment allocation and task order to ensure that the system operates stably and efficiently in different environments, thereby improving the system's adaptability and response speed.

[0020] In some embodiments, the execution method is as follows: Data Acquisition and Preprocessing: The data acquisition module obtains real-time data on equipment status, task progress, traffic flow, and environmental conditions within the port, utilizing various sensors to collect information. The collected data is transmitted to the optimization algorithm module via a standardized interface, where it undergoes noise reduction, formatting, and standardization to ensure data quality.

[0021] Global scheduling scheme generation: The GRASP algorithm in the optimization algorithm module begins working, searching for a high-quality initial scheduling scheme within a multi-dimensional parameter space by constructing a greedy random solution and conducting a local search. This scheme includes loading / unloading order, equipment allocation, and task priority, with the goal of minimizing total operation time and equipment energy consumption. The GRASP algorithm iteratively improves the scheduling scheme until it reaches the termination condition.

[0022] Local optimization and dynamic adjustment: Based on the initial scheduling scheme generated by GRASP, the SAC algorithm is used for local optimization. By constructing a deep neural network, the real-time equipment status and task progress of the port are analyzed, and task allocation and loading / unloading order are dynamically adjusted. For emergencies such as equipment failure and traffic congestion, the scheduling strategy is continuously optimized to improve the flexibility and adaptability of the scheduling scheme and ensure the efficient completion of scheduling tasks.

[0023] Task Execution and Monitoring: The task scheduling module translates the optimized scheduling plan into specific operation instructions, which are then sent to the equipment control system via standardized interfaces (such as RS485, Ethernet, Wi-Fi, and 5G). Each piece of equipment (such as quay cranes, yard cranes, and AGVs) receives the instructions and executes the container loading and unloading tasks. The real-time monitoring module continuously tracks task execution and equipment status to ensure the scheduling plan is implemented as scheduled and records task completion status.

[0024] Anomaly Handling and Feedback Optimization: When the status monitoring and real-time update module detects equipment failure, task delays, or other anomalies, it feeds the anomaly information back to the optimization algorithm module. Based on this anomaly feedback, the SAC algorithm adjusts the scheduling strategy in real time, reallocates tasks, and optimizes operation paths to ensure the continuity and efficiency of container loading and unloading tasks. The system dynamically adapts to the real-time changing environment of the port through this feedback mechanism.

[0025] Results Analysis and Adjustment: The strategy evaluation and feedback module analyzes the scheduling execution results, while the user interface module displays scheduling information and equipment status to operators. Operators can view the results and manually intervene or adjust scheduling parameters through the interface. Based on the feedback, the system further adjusts and optimizes the strategy, providing improvement references for subsequent scheduling tasks and ensuring the efficient operation and stability of the port scheduling system.

[0026] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) This invention achieves global and local dynamic optimization of port container scheduling by combining GRASP and SAC for collaborative optimization, quickly generates high-quality scheduling schemes and adjusts them in real time, significantly improves equipment utilization, shortens operation time, reduces energy consumption, and enhances the dynamic adaptability and robustness of scheduling strategies.

[0027] (2) The data acquisition module of the present invention integrates multiple sensors and data interfaces to acquire comprehensive data of port equipment and environment in real time, providing accurate input for optimization algorithms. Combined with data preprocessing, it improves the scientificity and reliability of scheduling decisions and ensures stable operation of the system in complex environments.

[0028] (3) The optimization algorithm module and the status monitoring module of the present invention work together to promptly detect and handle emergencies in port operations, such as equipment failures and task delays. By triggering re-optimization through a feedback mechanism, the stability and continuity of container loading and unloading tasks are ensured, and the system's intelligence and emergency response capabilities are improved.

[0029] (4) The real-time monitoring module and task scheduling module of the present invention work closely together to realize intelligent monitoring and dynamic optimization of the scheduling process. This ensures timely handling of abnormal situations, efficient execution of the scheduling plan, effective improvement of overall port operation efficiency, and guarantees the efficiency and continuity of port operations.

[0030] (5) The user interface module of this invention provides an intuitive and user-friendly graphical interface, displays scheduling information in real time, and supports operators to manually intervene, adjust scheduling parameters, and query historical data. This improves the system's operational flexibility and response speed, helps operators make scientific and efficient scheduling decisions, and ensures the stable and efficient execution of port container scheduling tasks.

[0031] (6) The present invention uses standardized interfaces to realize efficient data exchange and system integration between modules. The modular design of the system enables it to quickly adapt to the needs of different port environments, facilitates deployment and maintenance, and significantly improves the level of port management and intelligence. Attached Figure Description

[0032] Figure 1 This is a diagram showing the overall architecture of the port container scheduling system in this invention; Figure 2 This is a flowchart of the data acquisition module in this invention; Figure 3 This is a flowchart of the optimization algorithm module in this invention; Figure 4 This is a flowchart of the status monitoring module in this invention; Figure 5 This is the overall system operation flowchart in this invention (combining GRASP and SAC algorithms). Figure 6 This is a flowchart of the user interface module in this invention. Detailed Implementation

[0033] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0034] like Figure 1As shown, a port container scheduling system based on GRASP and SAC includes: a data acquisition module, an optimization algorithm module, a task scheduling module, a user interface module, and a real-time monitoring module.

[0035] The core of a port container scheduling system lies in the collaborative work between its various modules. The data acquisition module, located at the system's front end, is responsible for obtaining real-time data from the port's equipment and environmental monitoring systems to support scheduling decisions. The acquired data is transmitted to the optimization algorithm module via standardized interfaces (such as Ethernet, USB, RS485, etc.). The optimization algorithm module uses GRASP and SAC algorithms for global and local optimization to generate scheduling schemes. The task scheduling module then translates the optimization results into specific operational instructions and transmits them to the control systems of various equipment, such as quay cranes, yard cranes, and AGVs. The real-time monitoring module monitors the task execution process, detects equipment status and task progress, and feeds back any anomalies to the optimization algorithm module to ensure continuous optimization of the scheduling strategy and the stability of its execution. The user interface module provides operators with scheduling information display, equipment control, and anomaly alarm functions, supporting manual adjustment and real-time monitoring of scheduling tasks.

[0036] The data acquisition module includes equipment status sensors, intelligent load sensors, high-definition vision sensors, and environmental status sensors, integrating various sensors and data interfaces. Equipment status sensors (such as load sensors, position sensors, and equipment temperature sensors) monitor the workload and status of various equipment; task progress sensors (such as RFID readers, GPS modules, and yard cameras) track container movement and task completion; high-definition vision sensors (such as LiDAR, millimeter-wave radar, and traffic video surveillance equipment) acquire internal traffic conditions; and environmental status sensors (such as temperature, humidity, and wind speed sensors) collect port meteorological data. The data acquisition module transmits data to the optimization algorithm module via interfaces (such as RS232, I2C, CAN, and Ethernet) for scheduling optimization.

[0037] The optimization algorithm module is the core of this system, as shown in Figure 3. This module adopts a collaborative optimization method combining GRASP and SAC to achieve global and local optimization of port container scheduling.

[0038] First, the GRASP algorithm is used for global optimization, generating an initial scheduling scheme through a construction phase and a local search phase. In the construction phase, based on the problem's objective function and constraints, a greedy random strategy is used to select tasks and devices for allocation, forming an initial feasible solution. In the local search phase, a neighborhood search is performed on the initial solution, using operations such as swapping and insertion to find a better solution, ensuring the overall scheduling scheme achieves high quality. The optimization scheme considers multiple objectives, including minimizing total job time, reducing device energy consumption, and optimizing device utilization.

[0039] Building upon global optimization, the SAC algorithm employs deep neural network learning for local optimization, dynamically adjusting the execution strategy of loading and unloading tasks based on real-time environmental data. The SAC algorithm utilizes a flexible actor-critic architecture: the Actor network outputs a stochastic policy with entropy regularization, while the Critic network estimates the state-action value function. By maximizing cumulative reward and policy entropy, it achieves a balance between exploration and exploitation. The SAC algorithm is primarily responsible for dynamically adjusting based on data such as port equipment status, task progress, and traffic flow, optimizing task allocation and equipment operation, improving scheduling flexibility and adaptability, and rapidly responding to real-time environmental changes.

[0040] The optimization algorithm module forms a closed-loop feedback system through continuous data input, global optimization, local optimization, self-learning mechanisms, and feedback adjustments. This system continuously adjusts and optimizes scheduling strategies to cope with the complex and ever-changing port environment and dynamic task requirements. The module's design ensures the efficient operation and real-time response capability of the port scheduling system, thereby improving overall operational efficiency and stability.

[0041] The task scheduling module is the core of this system's execution, as shown in Figure 4. Based on the scheduling scheme generated by the optimization algorithm module, this module transforms the results of global and local optimizations into specific equipment instructions to execute actual container loading, unloading, and transportation operations. The task scheduling module first receives the optimal scheduling scheme generated by the GRASP and SAC algorithms. This scheme includes information such as the container loading / unloading sequence, equipment usage plan, and task allocation.

[0042] The task scheduling module ensures precise execution by refining these scheduling schemes into specific operational instructions. Specifically, the module is responsible for breaking down optimized loading and unloading sequences and equipment usage plans into instructions for various types of port equipment, including quay cranes, yard cranes, and automated guided vehicles (AGVs). As shown in Figure 4, these instructions are transmitted to the control systems of each piece of equipment through multiple communication methods: instructions for quay cranes and yard cranes are typically transmitted via RS485 and Ethernet interfaces to ensure stability and real-time performance; instructions for AGVs are transmitted at high speed via 5G networks to meet dynamically changing task requirements.

[0043] Upon receiving instructions, each equipment control system (such as quay cranes, yard cranes, and AGVs) executes container loading and unloading tasks according to the plan. The task scheduling module is not only responsible for issuing initial instructions but also continuously monitors the equipment's operational progress and task completion status, dynamically adjusting task allocation based on real-time feedback. Through a feedback mechanism, the task scheduling module can quickly adjust the equipment scheduling plan when it detects abnormal equipment status, task delays, or other unforeseen circumstances, ensuring the efficient execution and continuity of container loading and unloading tasks.

[0044] In quay crane operations, the task scheduling module can dynamically adjust the order of container loading and unloading and equipment priority to avoid equipment idleness and improve overall operational efficiency. For AGVs, the module can update their path planning and speed in real time to ensure optimal transportation routes and times. In yard crane operations, the task scheduling module dynamically adjusts the task allocation and movement paths of the yard cranes based on the container positions and loading / unloading task priorities within the yard.

[0045] The intelligent management of the task scheduling module not only ensures the coordination and efficient operation of various devices, but also achieves optimal execution of scheduled tasks through real-time adjustment and feedback mechanisms. This fully leverages the performance of port equipment and improves the overall efficiency and resource utilization of port container scheduling. Through this intelligent scheduling mechanism, the system can dynamically adapt to the complex and ever-changing environmental conditions and task requirements within the port, significantly improving scheduling efficiency and system stability.

[0046] The status monitoring module ensures stable task execution and dynamic adjustment, as shown in Figure 5. Through real-time data acquisition and intelligent analysis technology, it comprehensively monitors equipment operating status, task progress, traffic flow changes, and environmental parameter fluctuations during port operations. An anomaly detection model built using machine learning algorithms can quickly and accurately identify anomalies such as equipment failures, task delays, and traffic congestion. Once an anomaly is detected, a feedback mechanism is immediately activated, feeding the anomaly information and related data back to the intelligent decision optimization module, triggering a re-optimization process to ensure the real-time effectiveness of the scheduling plan and the continuity of operations. The status monitoring module works closely with the data acquisition module to ensure the timeliness and accuracy of information flow. Through a closed-loop control mechanism, it monitors the operating status of each piece of equipment and task progress, and updates the scheduling plan in real time. The module's design enables rapid identification and response to anomalies, effectively reducing the impact of delays and failures on the overall scheduling task, and ensuring the efficiency and continuity of port operations.

[0047] The user interface module provides port managers and operators with real-time operational information through an intuitive and user-friendly visual interface. This includes dynamic charts and map displays of key indicators such as equipment status, task progress, and resource consumption. It supports online adjustment of scheduling strategy parameters, such as task priority weights and equipment resource allocation ratios, enabling flexible manual intervention. As shown in Figure 6, this interface is used for real-time monitoring of system status and task progress, displaying key information such as equipment status, task execution, and port traffic flow. Operators can view current scheduling parameters through the interface and manually intervene when abnormalities are detected, such as triggering emergency equipment shutdown or rescheduling. The user interface module also supports real-time adjustment of scheduling parameters and provides historical data analysis functions, enabling operators to optimize scheduling strategies based on data feedback, improving system flexibility and responsiveness.

[0048] This module is designed as a multi-functional operating platform, featuring alarm notifications, equipment control, data querying, and log recording functions, supporting operators in comprehensive system management. Through the user interface module, operators can not only react quickly to and handle emergencies, but also make more scientific and efficient scheduling decisions based on historical data analysis, ensuring the stability and efficiency of port container scheduling tasks.

[0049] The intelligent decision optimization module integrates the advantages of GRASP and SAC algorithms. The GRASP algorithm utilizes heuristic rules to construct an initial solution space, rapidly generating multiple feasible initial scheduling schemes through a combination of random selection and a greedy strategy. Then, local search strategies, such as the 2-opt algorithm and simulated annealing, are applied to optimize the initial solutions, yielding a globally optimal initial scheduling scheme, providing a good starting point for subsequent optimizations. The SAC algorithm, based on Markov decision processes, models port scheduling tasks as an interaction between an agent and its environment. Through a flexible actor-critic architecture, the Actor network introduces entropy regularization when outputting scheduling strategies, increasing exploratory nature and avoiding getting trapped in local optima; the Critic network utilizes a double-Q network structure to more accurately estimate the state-action value function, improving learning efficiency. Based on real-time data, the SAC algorithm dynamically adjusts task allocation, equipment operation sequence, and resource allocation strategies, achieving fine-grained local optimization.

[0050] This invention models the port container scheduling problem as a multi-objective optimization problem, aiming to effectively utilize port resources, shorten container loading and unloading operation time, and reduce overall energy consumption. The model's optimization objectives include minimizing the total operating time and energy consumption of equipment, and maximizing equipment utilization, aiming to ensure the efficiency and economy of port operations through reasonable scheduling strategies.

[0051] During port operations, container loading and unloading tasks have different priorities and time constraints, and each task needs to be completed within a specific time window. These tasks are performed by various equipment such as quay cranes (QC), yard cranes (YC), automated guided vehicles (AGVs), and trucks. These devices have their own working capabilities, load capacities, and energy consumption characteristics. Therefore, the scheduling system needs to allocate tasks among different devices, taking into account the utilization efficiency and energy consumption of each type of equipment, to ensure the efficient completion of loading and unloading tasks.

[0052] Specifically, the port scheduling problem can be described as: within a given time window, allocating tasks to appropriate equipment to ensure that all tasks are completed efficiently according to priority and equipment performance characteristics. Due to the dynamic environment of ports, including equipment status changes and unforeseen events (such as equipment failures and weather changes), the scheduling system must be flexible and adaptable. To this end, this invention adopts a two-layer optimization structure, combining Greedy Random Adaptive Search (GRASP) and Flexible Actor-Critic Reinforcement Learning (SAC) to achieve the goals of global and local optimization.

[0053] The global optimization layer uses the GRASP algorithm to perform a comprehensive search within the multi-dimensional parameter space to generate an initial scheduling scheme. This algorithm ensures the discovery of high-quality initial solutions in complex multi-objective environments by constructing greedy random solutions and employing local search, balancing the allocation of tasks across different devices. The local optimization layer uses the SAC algorithm for real-time adjustments. SAC, through deep reinforcement learning, continuously learns and adapts to environmental changes (such as changes in equipment status and task progress) during port operations to optimally allocate equipment and adjust scheduling strategies, ensuring both flexibility and efficiency in scheduling.

[0054] The model's optimization objectives are threefold: minimizing total operation time (by reducing the total completion time of tasks to improve port operational efficiency); minimizing equipment energy consumption (by rationally allocating equipment and optimizing the operation sequence to reduce the total energy consumption of various types of equipment); and maximizing equipment utilization (by maximizing the ratio of actual working time of equipment to its total available time to improve equipment utilization efficiency).

[0055] In addition, the model needs to meet multiple constraints, including equipment capacity, task time window, priority, and internal port traffic flow, to ensure the feasibility and effectiveness of the model in practical applications.

[0056] The establishment of this mathematical model provides a solid theoretical foundation for subsequent algorithm implementation and guides the efficient operation of the scheduling system in complex and dynamic port environments. By using a two-layer optimization structure, this invention can not only find high-quality initial solutions for scheduling tasks but also flexibly adjust them according to actual operational conditions, improving the adaptability and robustness of the scheduling system in changing environments. The symbols, formulas, and solution methods in the model are described in detail below. 1. Problem Definition and Symbol Explanation Equipment collection: This indicates equipment such as quay cranes (QC), yard cranes (YC), and AGVs.

[0057] Task Collection: This indicates a container loading and unloading task to be scheduled.

[0058] Time window set: , representing a discrete time point within the scheduling period.

[0059] Decision variables: :Task Assign to device .

[0060] :Task The start and end times.

[0061] 1. Multi-objective optimization model Optimization goal: (1) Minimize the total delay time:

[0062] ( For the task (Latest completion time) (2) Minimize total energy consumption of the equipment:

[0063] ( For equipment Energy consumption per unit time (3) Maximize equipment utilization:

[0064] ( (Maximum available time for device j) Equipment capacity constraints:

[0065] Task time window constraints: , For the task Earliest start time Traffic flow constraints:

[0066] ( f kt Let k be the flow rate at time t. ) 2. GRASP Global Optimization Layer Construction Phase: Greedy random strategy: Select the optimal task-device allocation with probability α, and generate an initial solution set by random selection with a probability of 1-α.

[0067] Fitness function: Combined multi-objective weights:

[0068] in, This is the total delay time. The maximum allowable total delay time, Total energy consumption, Total available energy consumption Total active time, Total capacity time, These are the weighting coefficients.

[0069] Local search phase: 2-opt neighborhood operation: By randomly selecting and reversing a continuous sequence of tasks in the scheduling scheme, a new candidate scheduling scheme is generated, and its quality is evaluated by the fitness function. This process is continuously iterated and optimized to find a better global scheduling scheme.

[0070] 3. SAC Local Dynamic Optimization Layer Markov Decision Process (MDP) Modeling: State space S: Device load Task progress Traffic flow Environmental parameters (such as wind speed).

[0071] Action Space A: Task Reassignment Path adjustment .

[0072] Reward function design:

[0073] ( This represents the current total delay time. The delay threshold, For maximum permissible energy consumption, E real Actual energy consumption For device active time, Total available time, , , (Weighting coefficients) Strategy optimization: Actor Network: Outputs a Gaussian policy with entropy regularization. Enhance exploration capabilities:

[0074] in, It is a mean function. Let N be the state vector, and N be a multivariate normal distribution. Let Variance be the variance.

[0075] Critic Network: Double-Q network reduces estimation bias and updates the formula.

[0076] in, For the target Q value, For instant rewards, As a discount factor, For the Critic network, For the next state, For the next action, For Q network indexing.

[0077] 4. Collaborative optimization process GRASP generates initial solutions: It outputs a globally optimal solution through a greedy randomization strategy and a 2-opt search. .

[0078] SAC Dynamic Adjustment: Based on Real-Time Status Actor network generates actions Critic evaluation function, update strategy .

[0079] Closed-loop feedback: The real-time monitoring module detects anomalies (such as equipment failure) and triggers SAC re-optimization or GRASP global restart.

[0080] 5. Innovation and Differentiation GRASP vs QPSO: GRASP avoids premature convergence through a greedy stochastic strategy, while QPSO relies on particle swarm quantum behavior.

[0081] SAC vs DDPG: SAC uses entropy regularization and double-Q network to enhance exploration stability, while DDPG only uses deterministic policies.

[0082] Formula differentiation: fitness function The introduction of equipment idle rate index differs from the linear combination of energy consumption and time in QPSO.

[0083] SAC reward function It clearly distinguishes between real-time energy consumption and latency thresholds, unlike the single latency penalty of DDPG.

[0084] To solve the above mathematical model, this invention employs a two-layer optimization strategy. First, in the global optimization layer, the GRASP algorithm is used for a global search, generating an initial scheduling scheme through a construction phase and a local search phase, ensuring a high-quality initial solution is found. Then, in the local optimization layer, the SAC algorithm refines the initial scheme based on real-time data. The SAC algorithm learns real-time port status information through a deep neural network, dynamically adjusting task allocation and equipment operation to improve the adaptability and flexibility of the scheduling scheme. The specific operation steps are as follows: Step 1: Initialization and Data Preprocessing 1. Parameter Configuration Set the greedy coefficient of the GRASP algorithm Number of iterations for local search .

[0085] Configure the SAC algorithm's Actor / Critic network structure (3 fully connected layers, 256 hidden nodes), and set the learning rate. Discount factor Experience pool capacity

[0086] 2. Data Collection and Standardization The device status (location, load, energy consumption) is collected in real time using lidar, RFID, and temperature and humidity sensors. Task progress Traffic flow and environmental data (wind speed V) wind .

[0087] The original data is denoised, standardized (Z-score), and anomaly detected (based on the 3σ principle) to generate a standardized dataset.

[0088] Step 2: GRASP global optimization generates initial scheduling scheme 1. Construction Phase Task sorting: by task deadline Sort in ascending order to generate a task queue.

[0089] Greedy random allocation For each task Calculate the candidate device set

[0090] in, For the set of candidate devices, Remaining time for the device Let be the end time of task i. Let i be the start time of task i; With probability Choose the equipment with the lowest energy consumption. , by Random selection

[0091] Update device remaining time Generate initial solution set

[0092] 2. Local search phase Neighborhood operations: Randomly select 3 tasks , For device identifiers, exchange their device assignments.

[0093] Fitness evaluation: Calculate the fitness value of the new solution like Then accept the new solution and repeat the iteration. Next, output the globally optimal solution. .

[0094] Step 3: SAC Dynamic Local Optimization and Real-time Adjustment 1. Markov Decision Process Modeling State vector s:

[0095] Wind speed; Action vector a:

[0096] 2. Strategy Optimization Process Critic Network: A dual-Q network for evaluating state-action values and Update the target by taking the minimum value.

[0097] Actor Network: Input 8, output mean μ and variance o, generate Gaussian policy.

[0098] Reward calculation: calculated based on real-time data.

[0099] 3. Online learning and adjustment Environmental state 8 is sampled every 1 second, and actions are executed. Adjust task assignment or path If a device malfunction is detected (e.g.) This immediately interrupts the current task and triggers SAC reallocation.

[0100] Step 4: Exception Handling and Closed-Loop Feedback 1. Anomaly Detection Equipment malfunction: Vibration sensor data Task delay lasting 3 seconds: (default

[0101] 2. Feedback Mechanism If the anomaly level is Level 1 (single device failure), SAC will adjust the task allocation locally. If the anomaly level is Level 2 (multiple devices down or traffic paralysis), GRASP will be triggered for global restart optimization.

[0102] Step 5: Task execution and distribution Task execution and monitoring: The task scheduling module translates the optimized scheduling plan into specific operation instructions and sends them to the equipment control system via standardized interfaces (such as RS485, Ethernet, Wi-Fi, and 5G). Upon receiving the instructions, each piece of equipment (such as quay cranes, yard cranes, and AGVs) executes the container loading and unloading tasks according to the plan. The real-time monitoring module continuously monitors the task execution status and equipment status to ensure the scheduling plan is implemented as predetermined and records task completion information. Step 6: Data Interaction and Strategy Adjustment The user interface module provides operators with a display of scheduling information and a control interface, showing the current task status and equipment conditions. Operators can view the scheduling execution results in real time through the interface and manually intervene in abnormal situations or adjust scheduling parameters. Based on the scheduling execution results and operator feedback, the system further adjusts and optimizes its strategies, providing improvement references for subsequent scheduling tasks. Through continuous monitoring and feedback adjustments, the efficient operation and stability of the port scheduling system are ensured.

[0103] Through this dual-layer optimization structure, the present invention can achieve efficient scheduling optimization at both the global and local levels, ensuring maximum utilization of port equipment, minimum operation time, and minimum energy consumption, thereby improving port operation efficiency and management level.

[0104] This invention provides a port container scheduling system and its operation method based on GRASP and SAC collaborative optimization, as shown in Figures 1 to 6. The specific implementation steps are as follows: Step 1: Data Acquisition and Preprocessing. As shown in Figure 2, the data acquisition module obtains real-time data on equipment status, task progress, and traffic flow within the port, utilizing various sensors including load sensors, GPS modules, and traffic monitoring equipment. The acquired data is transmitted to the optimization algorithm module via standardized interfaces such as RS485, Ethernet, and 5G. The data acquisition module performs preliminary data processing, including noise reduction, formatting, and standardization, ensuring data accuracy and real-time performance, providing high-quality input for subsequent optimization processes.

[0105] Step 2: Global Scheduling Scheme Generation. As shown in Figure 3, the optimization algorithm module first uses the Greedy Random Adaptive Search (GRASP) algorithm for global optimization. By constructing a greedy random solution and performing a local search, a high-quality initial scheduling scheme is searched in the multi-dimensional parameter space. This scheme includes loading / unloading order, equipment allocation, and task priority, with the goal of minimizing total operation time and equipment energy consumption. GRASP iteratively improves the scheduling scheme until the termination condition is met.

[0106] Step 3: Local Optimization and Dynamic Adjustment. Based on the initial scheduling scheme generated by GRASP, as shown in Figure 3, the system further uses the Flexible Actor-Critic (SAC) reinforcement learning algorithm for local optimization. SAC constructs a deep neural network to analyze the current equipment status and task progress of the port in real time, dynamically adjusting task allocation and loading / unloading order to improve the flexibility and adaptability of the scheduling scheme. With the support of real-time data, the SAC algorithm continuously optimizes the scheduling strategy for unexpected situations such as equipment failure and traffic congestion, ensuring the efficient completion of scheduling tasks.

[0107] Step 4: Task Execution and Monitoring. As shown in Figure 4, the task scheduling module translates the optimized scheduling plan into specific operation instructions and sends them to the equipment control system via standardized interfaces (such as RS485, Ethernet, Wi-Fi, and 5G). Upon receiving the instructions, each piece of equipment (such as quay cranes, yard cranes, and AGVs) executes the container loading and unloading tasks according to the plan. The real-time monitoring module continuously monitors the task execution status and equipment status to ensure that the scheduling plan is implemented as predetermined and records the task completion status.

[0108] Step 5: Anomaly Handling and Feedback Optimization. As shown in Figure 5, when the status monitoring module detects equipment failure, task delay, or other anomalies, it feeds the anomaly information back to the optimization algorithm module. Based on this anomaly feedback, the SAC algorithm adjusts the scheduling strategy in real time, reallocates tasks, and optimizes operation paths to ensure the continuity and efficiency of container loading and unloading tasks. The system dynamically adjusts its strategy through a feedback mechanism to adapt to the real-time changing environment of the port.

[0109] Step 6: Result Analysis and Adjustment. As shown in Figure 6, the user interface module provides a scheduling information display and control operation interface, showing the current task status and equipment conditions. Operators can view the scheduling execution results in real time through the interface and manually intervene in abnormal situations or adjust scheduling parameters. Based on the scheduling execution results and operator feedback, the system further adjusts and optimizes strategies, providing improvement references for subsequent scheduling tasks. Through continuous monitoring and feedback adjustments, the efficient operation and stability of the port scheduling system are ensured.

[0110] This invention combines GRASP with SAC (such as...) Figure 1 As shown in the diagram, the complete system architecture is illustrated, with each module having a clear division of labor and working closely together. The data acquisition module provides basic data for other modules, the optimization algorithm module optimizes the scheduling scheme based on this data, the task scheduling module executes the optimization scheme, the status monitoring module provides feedback to adjust the strategy, and the user interface module enables user interaction with the system. This architectural design enables the system to form a highly efficient and collaborative whole, improving the overall performance and stability of the scheduling system.

[0111] The data acquisition module of this invention integrates multiple sensors and data interfaces (such as...) Figure 2 As shown, numerous types of sensors ensure comprehensive collection of various port data, while multi-channel data processing units preprocess the data to guarantee data quality. Multiple standardized data interfaces enable fast and stable data transmission, providing reliable and timely data support for subsequent optimization algorithms, thus enabling more accurate scheduling decisions.

[0112] The intelligent decision optimization module and the status monitoring module of this invention work together (e.g. Figure 5 As shown in the diagram, the GRASP algorithm performs a global search to generate an initial scheduling scheme, while the SAC algorithm performs local optimization based on real-time data. The combination of these two algorithms optimizes the scheduling strategy at different levels, considering both overall resource allocation and flexible responses to real-time changes at the port. This improves the quality of the scheduling scheme and its adaptability to complex environments, effectively enhancing port operational efficiency.

[0113] The status monitoring module and task scheduling module of this invention work closely together, collecting data in real time through multiple sensors and using machine learning algorithms to build anomaly detection models for rapid and accurate anomaly identification. Different anomaly levels correspond to different handling methods, allowing for timely adjustments to scheduling strategies to reduce the impact of anomalies on port operations and ensure operational continuity and stability.

[0114] The user interface module of this invention provides an intuitive and user-friendly graphical interface (such as...). Figure 6The diagram illustrates the overall system operation flow. From data input preprocessing to the collaborative optimization of GRASP and SAC algorithms, and then to task execution, anomaly handling, and feedback adjustments, a closed-loop operation mechanism is formed. This ensures that the system continuously optimizes in the ever-changing port environment, maintains efficient operation, and improves the overall port operation and management level.

[0115] This invention uses standardized interfaces to achieve efficient data exchange and system integration between modules. The modular design of the system enables it to quickly adapt to the needs of different port environments, facilitates deployment and maintenance, and significantly improves the level of port management and intelligence.

Claims

1. A port container scheduling system based on GRASP and SAC, characterized in that, include: The data acquisition module is used to collect real-time data from port equipment and environmental sensors and to preprocess it. The optimization algorithm module is used to receive data output from the data acquisition module and generate a scheduling scheme, or to receive abnormal data fed back by the real-time monitoring module, automatically start the global optimization process of the GRASP algorithm and / or the local optimization process of the SAC algorithm, and regenerate the scheduling scheme. The task scheduling module is used to receive scheduling plans and execute and adjust specific task allocation and equipment scheduling operations; The real-time monitoring module is used to monitor the device status and task execution, and transmit abnormal information to the optimization algorithm module.

2. The port container scheduling system based on GRASP and SAC according to claim 1, characterized in that, The data acquisition module includes: Position sensors are used to monitor the position and trajectory of equipment and containers; Load sensors are used to monitor equipment load and operating status; Environmental sensors are used to acquire port meteorological data; Traffic flow sensors are used for real-time data acquisition, predictive analysis, and closed-loop feedback. Equipment status sensors are used to monitor the operating status of critical equipment; A multi-channel data processing unit that supports multiple data interfaces is used to perform noise filtering, standardization, and anomaly detection on the collected data to ensure data quality and provide reliable input for the optimization algorithm module.

3. The port container scheduling system based on GRASP and SAC according to claim 1, characterized in that, In step (1), the optimization algorithm module includes a GRASP algorithm submodule for global optimization and a SAC algorithm submodule for local dynamic optimization; The GRASP submodule performs a global search in the multidimensional parameter space, and generates an initial scheduling scheme by constructing a greedy random solution and a local search, taking into account factors such as task priority, equipment allocation and loading / unloading order; The SAC submodule constructs a deep neural network optimization strategy based on the initial scheduling scheme generated by the GRASP submodule. Its state space includes equipment load, task progress, traffic flow and environmental state information, and its action space covers equipment task allocation and priority adjustment operations. It uses real-time data to dynamically adjust the local scheduling strategy and combines the Actor-Critic structure to optimize the scheduling scheme, thereby improving scheduling flexibility and adaptability.

4. The port container scheduling system based on GRASP and SAC according to claim 1, characterized in that, The global optimization process of the GRASP algorithm includes: During the initialization phase, parameters are set, including the random factor used in the construction phase and the neighborhood search radius for local searches. During the construction phase, based on the port equipment's processing capacity and task time requirements, a combination of tasks and equipment is randomly selected to form an initial feasible solution; During the local search phase, the 2-opt method is used to optimize the initial solution, and the quality of the solution is evaluated by calculating the fitness function, which is defined as: in, For the task The actual completion time, For the task Maximum allowed completion time For equipment Energy consumption per unit time For equipment Homework time, For equipment Total available time, For the task The delay time, For the task Maximum allowable delay time These are the weighting coefficients; When the maximum number of iterations is reached or the fitness value converges, the initial scheduling scheme is output.

5. The port container scheduling system based on GRASP and SAC according to claim 4, characterized in that, The local optimization process of the SAC algorithm includes: Port scheduling tasks are modeled as Markov decision processes. In model construction and parameter definition, state space parameters are determined based on remaining equipment capacity, task urgency, and time windows. Action policies in the action space aim to maximize long-term cumulative rewards and policy entropy. In the training and updating of the deep reinforcement learning network, the Actor network outputs an entropy-regularized stochastic scheduling policy based on the current state. Critic network estimates state-action value function ; The reward function is defined as: in, These are the weighting coefficients; At each time step, the system samples the current state, the Actor network outputs an action, and after execution, obtains the new state and immediate reward. The parameters of the Actor and Critic networks are updated using gradient descent. When the maximum number of iterations or the reward function convergence condition is met, the optimized local scheduling scheme is output.

6. The port container scheduling system based on GRASP and SAC according to claim 1, characterized in that, The task scheduling module includes: The task parsing unit is used to receive the scheduling scheme from the optimization algorithm module and parse it into a sequence of device operation tasks; The instruction generation unit is used to generate a set of device control instructions based on the parsing results, including operation steps, execution order, time and device parameters; The communication control unit is used to transmit instructions to the equipment control system through various communication protocols, and to receive equipment execution feedback in real time. Based on the feedback, it adjusts task allocation to ensure efficient task execution.

7. The port container scheduling system based on GRASP and SAC according to claim 1, characterized in that, The real-time monitoring module includes: The equipment status monitoring unit is used to monitor the operating status of equipment using sensors such as vibration, temperature, and current. The task progress tracking unit is used to record and analyze the task execution progress in real time. The anomaly handling unit is used to generate an anomaly report and feed it back to the optimization algorithm module when equipment failure, task delay or traffic congestion is detected, triggering the optimization process, adjusting the scheduling strategy, and ensuring stable system operation.

8. The port container scheduling system based on GRASP and SAC according to claim 1, characterized in that, It also includes a user interface module, which includes: The data display unit is used to display the port equipment status, task execution status, scheduling plan and energy consumption data in real time. The control panel unit provides an operation interface, allowing users to adjust scheduling parameters and modify scheduling schemes. The emergency intervention unit is used to provide emergency operation options, such as emergency pause and equipment switching, in the event of system malfunction, thereby improving the system's emergency response capabilities.

9. The operation method of the port container scheduling system based on GRASP and SAC according to any one of claims 1-8, characterized in that, Includes the following steps: The data acquisition module collects port equipment, task, traffic and environmental data in real time, performs preprocessing such as noise reduction and standardization, and transmits the processed data to the optimization algorithm module through a standardized interface; The GRASP submodule in the optimization algorithm module analyzes the preprocessed data, searches in the multidimensional parameter space, and generates an initial scheduling scheme. The goal is to optimize the loading and unloading sequence, equipment allocation, and task priority, thereby reducing the total operation time, equipment energy consumption, and task delay rate. The SAC submodule in the optimization algorithm module is based on the initial scheduling scheme generated by the GRASP submodule. It combines real-time data to build a deep neural network, dynamically adjusts task allocation and loading / unloading order, responds to port emergencies and equipment status changes, and improves the flexibility and adaptability of the scheduling scheme. The task scheduling module transforms the optimized scheduling scheme into operation instructions and sends them to the equipment control system. The real-time monitoring module monitors task execution and equipment status to ensure that the scheduling scheme is implemented as planned. When the real-time monitoring module detects an anomaly, it feeds back the anomaly information to the optimization algorithm module, and adjusts the scheduling strategy in real time based on the feedback to ensure that container loading and unloading tasks are carried out continuously and efficiently. The user interface module displays scheduling information to operators, who can manually intervene or adjust scheduling parameters based on execution results and actual needs. The system further optimizes strategies based on feedback, providing a reference for subsequent scheduling tasks.

10. The operating method according to claim 9, characterized in that, When the real-time monitoring module detects an anomaly, it feeds back the anomaly information to the optimization algorithm module, and adjusts the scheduling strategy in real time based on the feedback, including: The real-time monitoring module continuously monitors equipment status, task progress, and traffic flow. When an anomaly occurs, the feedback mechanism transmits the anomaly information to the optimization algorithm module. Based on the feedback data, the optimization algorithm module automatically initiates the GRASP algorithm global optimization or the SAC algorithm local optimization process, recalculates the scheduling strategy, and adjusts equipment allocation and task order to ensure that the system operates stably and efficiently in different environments, thereby improving the system's adaptability and response speed.

Images