An automated loading operation scheduling method for a container terminal

Abstract

The application discloses an automatic loading operation scheduling method for a container wharf, which is realized through the following steps: determining a candidate container list that can be operated currently based on a preset ship position dependence relationship and a real-time yard state through feasibility screening; respectively performing cost evaluation and opportunity utility evaluation on the candidate operation, the former quantifying the cost of equipment movement, turning operation, etc., and the latter quantifying the benefits brought by unlocking key operation, forming double-box cooperation or matching wharf resources in a forward-looking manner; calculating a final decision score through a weighted algorithm combining utility and cost, and selecting an optimal operation to generate an instruction accordingly. The application decomposes a complex scheduling problem into multiple evaluation levels, realizes fine and intelligent management of loading operation decision-making, and can effectively improve the overall efficiency, balance and economy of the automatic loading operation of the wharf.

Description

Technical Field

[0001] This invention relates to the field of automated port technology, and in particular to an automated ship loading operation scheduling method for container terminals. Background Technology

[0002] In modern container terminals, loading operations are the core link in the entire port logistics chain, and their efficiency directly affects the turnaround speed of ships and the overall throughput capacity of the terminal. Traditional loading operation scheduling usually relies on manual or semi-automated methods, assigning container loading tasks one by one according to a pre-established static stowage plan.

[0003] However, this approach faces numerous challenges in complex real-world operational environments. On one hand, terminal yards are dynamic environments where the physical location and status of containers, as well as the real-time location of operating equipment, are constantly changing. Static planning struggles to adapt to these changes, often leading to operational conflicts and efficiency bottlenecks. For example, if a planned container cannot be immediately retrieved because it is pinned down by other containers, a lack of intelligent system adjustment results in wasted time. On the other hand, simply following a fixed loading sequence often overlooks the actual costs incurred in executing specific tasks, such as the extensive movement of yard cranes, numerous container handling operations, and localized congestion or resource idleness caused by uneven command distribution. These factors collectively result in significant limitations of existing technologies in improving the overall smoothness, balance, and economy of loading operations.

[0004] How to combine complex loading requirements with dynamic terminal conditions and proactively assess the potential costs and future benefits of each operational decision in order to make globally optimal scheduling decisions while satisfying all constraints is a technical challenge that urgently needs to be solved in this field. Summary of the Invention

[0005] To overcome the shortcomings of the existing technology, this invention provides an automated ship loading operation scheduling method for container terminals. This method constructs a three-layer evaluation model, decomposing the complex scheduling problem into three logical levels: feasibility screening, cost assessment, and opportunity utility assessment. This achieves refined and intelligent management of ship loading operation decisions, specifically including the following steps:

[0006] S1. Feasibility screening: Based on the preset ship location dependency model and real-time status data of the terminal yard, select containers that are physically and logically feasible for operation from all containers to be loaded, forming a candidate operation list.

[0007] S2. Execution cost assessment: For each candidate task in the candidate task list, a quantitative assessment is performed from at least one preset cost dimension to determine its execution cost score; the cost dimension includes at least: equipment relocation cost, tamping operation cost, or potential conflict cost.

[0008] S3. Opportunity utility assessment: For each candidate operation in the candidate operation list, a forward-looking quantitative assessment is performed from at least one preset utility dimension to determine its opportunity utility score; the utility dimension includes at least: unlocking utility of unlocking subsequent key operations, synergistic utility of forming a double container operation, or resource matching utility of matching with terminal resources.

[0009] S4. Final decision and instruction generation: Based on the execution cost score and the opportunity utility score, calculate the final decision score for each candidate operation; based on the final decision score, select the optimal operation from the candidate operation list and generate the corresponding loading operation instruction.

[0010] Furthermore, the feasibility screening method in S1 includes:

[0011] The pre-set loading plan is transformed into a directed acyclic graph to represent the shipboard loading support relationship between containers;

[0012] It also obtains in real time the coordinates of each container's location in the terminal yard, its physical status, and the information on the container's overhead pressure points.

[0013] Furthermore, the specific steps of feasibility screening in S1 are as follows:

[0014] S11. Based on the directed acyclic graph, select containers whose parent node containers have been loaded from all containers to be loaded, and form a list of feasible positions on the ship.

[0015] S12. Based on the real-time status data of the terminal yard, further select containers whose physical status is operable from the list of feasible ship locations to form the final candidate operation list;

[0016] The conditions for the physical state to be operable are: there are no other containers above the container, it is not locked by any operation process in the system, and it is physically intact.

[0017] Furthermore, the execution cost score in S2 is calculated by incorporating the following costs:

[0018] The cost of moving equipment is proportional to the distance or time it takes for the yard crane responsible for performing the operation to move from its current real-time location to the location of the candidate container in the yard.

[0019] The cost of tamping operations is proportional to the number of other containers that need to be temporarily moved and placed on top of the candidate container to be retrieved.

[0020] Furthermore, when calculating the cost of the tamping operation, differentiated cost weights are set for different types of containers being tamped, including at least ordinary containers, refrigerated containers, or dangerous goods containers.

[0021] Furthermore, the potential conflict cost in S2 is proportional to the busyness of the operating area or equipment associated with the candidate container. The quantitative indicators of the busyness include the current number of orders for the yard bay where the candidate container is located, or the current number of orders for the quay crane responsible for it.

[0022] Furthermore, the "unlocking utility of unlocking subsequent key operations" in S3 is quantified as follows:

[0023] The assessment and statistics determine how many containers that were previously blocked or obstructed by the removal of current candidate containers, and which belong to higher priority operations in the preset loading sequence, can be transformed into a state that can be directly extracted, and are quantified based on the number or priority of the key containers that are unlocked.

[0024] Furthermore, the quantification method for "forming a synergistic effect of dual-box operation" in S3 is as follows:

[0025] The system detects whether the destination location of the candidate container is adjacent to or close to the destination location of another container that is already loaded or about to be loaded on the truck, either on the ship or in the yard. If the preset proximity condition is met, a higher cooperative utility score is assigned.

[0026] Furthermore, the "resource matching utility that matches the port resources" in S3 is quantified as follows:

[0027] Assess the degree of matching between the operational requirements of the candidate containers (such as the target quay crane) and the available terminal resources (such as the commanded load of each quay crane and the number of configured trucks); when the operation is assigned to a quay crane with a lower load or more abundant resources, its resource matching utility score is higher.

[0028] Furthermore, the final decision-making and instruction generation steps in S4 specifically include:

[0029] S41. Using a preset weighted algorithm, the execution cost score and the opportunity utility score are combined to calculate the final decision score for each candidate task. The calculation method can be expressed as follows:

[0030] Final decision score = α * Opportunity utility score - β * Execution cost score

[0031] Where α and β are preset weighting coefficients, and both are positive values;

[0032] S42. Select the job with the highest final decision score as the optimal job and generate instructions.

[0033] Beneficial effects:

[0034] This invention establishes a feasibility screening step, combining pre-defined ship location dependencies with the real-time physical state of the terminal yard to dynamically generate a list of candidate containers currently available for operation. This step ensures that subsequent decisions are based on physically and logically feasible operational options, thereby avoiding operational interruptions or conflicts caused by attempting to execute infeasible instructions, and improving the continuity and smoothness of the loading operation process.

[0035] This invention employs two independent evaluation steps—cost assessment and opportunity utility assessment—to conduct multi-dimensional quantitative analysis of scheduling decisions. In cost assessment, the method quantifies the direct costs required to execute different tasks, including but not limited to the cost of moving equipment from its current location to the target container location, and the cost of tamping operations necessary to retrieve the target container. In opportunity utility assessment, the method proactively quantifies the positive impact of executing a task on the overall loading schedule. For example, it prioritizes tasks that relieve constraints on the location of subsequent critical containers, or improves the efficiency of a single task by combining two adjacent containers to form a double-container operation.

[0036] This invention employs a final decision-making step, based on a pre-defined fusion algorithm, to comprehensively calculate the aforementioned cost and utility scores to select the optimal work instruction. This decision-making mechanism enables the system to choose the work option that yields higher overall benefits with lower execution costs from multiple feasible options. Compared to traditional methods that execute work in a fixed sequence, this invention, through its multi-level and multi-dimensional quantitative evaluation and decision-making process, can more rationally allocate terminal resources, reduce unnecessary equipment movement and handling operations, thereby achieving significant technical effects in improving the overall efficiency, balance, and economy of automated terminal ship loading operations. Attached Figure Description

[0037] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below with reference to a specific embodiment.

[0039] Example 1:

[0040] This embodiment provides an automated container terminal loading operation scheduling method based on hierarchical utility evaluation for guiding loading operations. Before executing this method, the scheduling system needs to complete data initialization. This initialization process includes: reading the global ship stowage plan and creating a data object for each container to be loaded, which contains information describing its physical support dependencies on the ship, thus logically forming a dependency network. Simultaneously, the system obtains and maintains the precise location information of all containers in the yard and the real-time dynamics of all available operating equipment through a real-time interface. The specific scheduling method includes the following steps:

[0041] S1: Feasibility Screening

[0042] The purpose of this step is to accurately identify and compile all immediately executable tasks from the complex dock environment, forming a dynamically updated "candidate task list." This step is the cornerstone of all subsequent evaluations and decisions. Through a rigorous two-stage filtering process, it ensures that every option in the list is logically and physically feasible.

[0043] Phase 1: Onboard location dependency check (logical feasibility filtering)

[0044] This method first performs a logical dependency check on the ship's position for each container in the global list of containers to be loaded. In the previous initialization steps, the global stowage plan has been transformed into a directed acyclic graph to accurately represent the physical support relationships between containers within the ship's hold. The specific check method is as follows:

[0045] For each container to be loaded, the method iterates through it and checks whether all its parent containers (i.e., all containers that must be loaded before it and serve as its physical support) in the dependency graph have been loaded. Only when all the parent containers of a container to be loaded have been loaded and in place is the method determined that it logically "meets the loading conditions" and added to a temporary "logically feasible list" for the next step of physical status check.

[0046] Phase Two: Physical Condition Inspection of the Stockyard (Physical Feasibility Filtering)

[0047] The scope of inspection at this stage will be temporarily expanded. This method will inspect the physical operability of all containers located on the top layer of the yard (regardless of whether they are on the "logically feasible list"). This aims to include top-level obstruction containers that, while not meeting the loading logic, may need to be removed as part of a "tampering" operation. The inspection criteria at this stage include:

[0048] Location accessibility: Confirm that the container is located on the top physical layer of its stack, with no other containers above it.

[0049] Status Availability: Confirm that the container has not been marked as logically locked by any other process (such as being damaged and awaiting inspection).

[0050] Based on the checks in the two stages described above, a corresponding list is generated. The list generation rules are as follows:

[0051] For containers in the "logically feasible list", if they pass the second-stage physical inspection, they are added to the final "candidate operation list" as a "shipping candidate operation".

[0052] For containers that are not in the "logically feasible list" (e.g., they have a parent node that has not been loaded onto the ship, or they are not in the loading plan at all), if they pass the physical inspection in the second stage, they are added to the final "candidate operation list" as a "tampering candidate operation".

[0053] To illustrate this screening process more clearly, let's take a specific scenario as an example:

[0054] In the loading plan, containers A, B, and C are to be loaded. The loading dependencies on the ship are as follows: A must be loaded before B (A is the parent node of B). C has no dependency on A or B.

[0055] The initial yard conditions are as follows: container A is located on the top layer of stack 1; container B is located on the top layer of stack 2; container C is pressed down by a barrier container D that is not in the loading plan; barrier container D is located on the top layer of stack 3.

[0056] The S1 screening process in the first round of decision-making cycle:

[0057] Perform the first phase of logical checks. Containers A and C, which have no unloaded parent nodes, pass the check. Container B, whose parent node A has not yet been loaded, does not satisfy the logical dependency and is filtered out. After this phase, the temporary "logically feasible list" includes A and C.

[0058] The second phase of physical checks is performed. In the "Logically Feasible List," A is at the top physical level and, after checking, is added to the final "Candidate Operation List" as a "Candidate Loading Operation." C is filtered out because it is not at the top physical level. Meanwhile, the check reveals that obstacle box D, although not in the loading plan, is at the top physical level, and therefore it is added to the list as a "Candidate Turnover Operation."

[0059] Final results of this round of screening: The current "candidate operation list" contains {ship loading operation A, tamping operation D}.

[0060] S2: Implementation Cost Assessment

[0061] This step calculates a quantified execution cost score for each task in the "Candidate Task List". This score is a numerical representation of the total cost required to execute the task, and is a weighted average of multiple cost sub-items.

[0062] One sub-item is equipment relocation cost, which is proportional to the physical distance the yard crane responsible for the operation needs to move from its current real-time location to the location of the candidate container, quantifying the basic time and energy consumption. Another sub-item is potential conflict cost, which consists of two parts: the total number of work orders already assigned in the yard bay where the candidate container is located, and the length of the queue of pending orders for the target quay crane for that container. The higher the sum of these two values, the greater the risk of congestion and potential waiting time in the work area, and the higher the cost.

[0063] In addition, this step includes a crucial sub-item: the estimation of future tamping costs. This cost is calculated by assessing which containers underneath will be exposed if the current candidate container is removed, and then calculating the difference in the number of tamping operations required for each of these exposed containers before and after the removal of the current candidate container. Summing these differences yields a quantified value representing the change in future tamping workload caused by the current operation, which serves as the cost figure.

[0064] Following the scenario in step S1, assuming that loading operation A has been completed and left the site, the list updated according to S1 is {loading operation B, turning operation D}. In the specific scenario, container B is located on the top layer of stack 2, and obstruction container D is located on the top layer of stack 3, with container C to be loaded pressing down on it.

[0065] (1) Assess the execution cost of "Loading Operation B":

[0066] Equipment relocation cost: Calculate the physical distance required for the yard crane responsible for the operation to move from its current location to stack 2 where container B is located. The greater the distance, the higher the score for this cost.

[0067] Turning and tamping operation cost: Since container B is already on the top layer of the stack, retrieving it does not require moving any other containers. Therefore, the direct turning and tamping cost incurred by performing this operation is 0.

[0068] Potential conflict cost: Assess the busyness of the target operating area (the bay where stack 2 is located) and the target quay crane for container B. The system will count the number of existing operation orders in the area and the length of the order queue for the target quay crane. The higher the total number, the greater the risk of conflict and waiting, and the higher the cost score for this item.

[0069] (2) Assess the execution cost of "Tamping Operation D":

[0070] Equipment relocation cost: Similarly, calculate the physical distance required for the field bridge to move from its current position to the stack 3 where the obstacle box D is located, and convert it into a cost score.

[0071] Turnover operation cost: Since operation D itself is a turnover operation (moving an obstacle container), its direct "turnover operation cost" is the movement itself. More importantly, this method will assess the "future" impact of moving container D. In this scenario, moving D is for the purpose of retrieving C. This cost assessment will focus on whether the ease of retrieving C changes after moving D, and whether it will create new turnover requirements in other locations in the yard.

[0072] Potential congestion costs: Assess the congestion level of the stack 3 area where barrier box D is located. Similar to the assessment of B, consider the instruction density and resource usage of this area to quantify its potential congestion costs.

[0073] Following the evaluation in S2, both Operation B and Operation D will receive a quantified, comprehensive execution cost score. This score reflects the "cost" of executing each operation; for example, if container B is closer to the yard crane and its target quay crane is available, its execution cost may be lower than that of container D, and vice versa. This cost score will be combined with the opportunity utility score calculated in the next step, S3, for the final decision.

[0074] S3: Opportunity Utility Assessment

[0075] This step is performed in parallel with the cost assessment, calculating an opportunity utility score for each task in the "candidate task list". This score is a quantitative assessment of the forward-looking and strategic benefits that performing the task may bring, and is also a weighted sum of multiple utility sub-items.

[0076] One utility sub-item is collaborative operation utility. This method will examine whether candidate operations have the potential to form a "two-container collaborative operation". If a candidate container (such as a standard 20-foot container) is found to have another candidate container that can be paired with it and carried out in a single double-lift operation near its target location in the yard and ship hold, the candidate operation is assigned a significantly high score to incentivize the prioritization of such efficient operation modes.

[0077] Another utility sub-item is resource matching utility, which measures the degree of immediate matching between a job and available resources. Its value can be directly taken as the number of trucks currently idle and waiting under the target quay crane for the job. The more trucks waiting, the more immediately the job can be responded to and executed, the higher the resource utilization efficiency, and therefore the higher its utility score.

[0078] Another utility sub-item is the critical path unlocking utility, which is the core mechanism of this method to break operational deadlocks and ensure process continuity. It is calculated as follows: determine which containers directly blocked by the current candidate operation will unlock (i.e., make it physically accessible as the top-level container) by executing the current candidate operation; then sum the preset priority values ​​of all these unlocked containers. This sum is the utility score, which quantifies the strategic value of the current operation in advancing the overall loading plan.

[0079] Following the scenario listed in S2, the current candidate operation list is {loading operation B, tamping operation D}. At this time, the yard status is: container A has been loaded; container B is the top-level container and can be processed directly; container D is also the top-level container, but it is pressing down on container C, which is waiting to be loaded. In this scenario, it is assumed that the priority of the containers waiting to be loaded is A=10, B=9, and C=8.

[0080] (1) Evaluate the opportunity utility of “loading operation B”:

[0081] Collaborative Operation Utility: Assuming container B is a standard 40-foot container, it does not have the potential to form a "dual-container collaborative operation" with other containers. Therefore, this utility score is 0.

[0082] Resource matching utility: This evaluates the number of trucks waiting idly under the target quay crane assigned to B. If there is currently one truck waiting, the utility score is 1. The more trucks waiting, the higher the score.

[0083] Critical path unlocking utility: Since container B is the top-level container, removing it will not unlock any containers that are blocked by it and are in the loading schedule. Therefore, this utility is 0.

[0084] (2) Evaluate the opportunity utility of “Turning and Turning Operation D”:

[0085] Collaborative operation utility: Operation D is a tamping operation, which does not directly participate in loading onto the ship, and is usually not evaluated. This utility is 0.

[0086] Resource matching utility: Turnover operations are typically completed within the yard and are not directly matched with quay cranes and container trucks. Therefore, this utility is 0.

[0087] Unlocking the critical path utility: This is the core of evaluating the value of Operation D. Performing "Tampering Operation D" directly unlocks (makes the top-level container physically accessible) the container C that it was pressing down. According to the calculation rules, the preset priority values ​​of all unlocked containers are summed. In this scenario, only container C is unlocked; therefore, the "Unlocking the critical path utility" score for Operation D = the priority of container C = 8.

[0088] In addition to the scenarios mentioned above, there may be instances where no containers for loading operations are currently listed. In such cases, the tamping operation will be evaluated directly. To illustrate this more clearly, the following scenario assumptions are made:

[0089] According to the loading plan, the containers most urgently needed for loading are A, B, and C, with priority values ​​of 10, 9, and 8 respectively. However, in the yard, the high-priority containers A and B are both blocked by a top-level container F, which is neither awaiting loading nor urgently needed for loading; container C is blocked by a top-level container G. At this point, after the screening in step S1, the truly feasible "candidate job list" only includes jobs F and G.

[0090] Although loading operations are not performed at this time, the utility calculations for the containers in the list need to be performed, specifically including:

[0091] Evaluation of task F: Executing the task "Move box F" will simultaneously unlock (expose) boxes A and B below it. This method then sums the priority values ​​of boxes A and B to obtain the "unlock critical path utility" score for task F: 10 + 9 = 19 points.

[0092] Evaluation of task G: Performing the task "Move box G" will only unlock box C below it. Therefore, the "unlock critical path utility" score of task G is the priority value of box C: 8 points.

[0093] Although both moving box F and moving box G are currently considered "tampering" tasks, task F has a much higher "critical path unlocking utility" (19 points vs. 8 points) than task G. Therefore, in the final decision of S4, this method will most likely select F as the optimal task and generate instructions.

[0094] Through the above quantitative assessment, this method, when there are no directly selectable loading tasks, does not stagnate but identifies the most strategically valuable solution—the repackaging operation. After executing the "move container F" instruction, the yard status is refreshed. In the next decision-making cycle, the high-priority containers A and B will become "top-level" operable containers, thus smoothly entering the "candidate operation list" and ensuring the continuation of the entire loading process.

[0095] S4: Final Decision and Instruction Generation

[0096] This step marks the decision endpoint of this scheduling method. The method substitutes the execution cost score and opportunity utility score obtained by each candidate job in the first two steps into a pre-defined utility-cost fusion function. The function multiplies the opportunity utility score by a benefit weighting coefficient and subtracts the execution cost score multiplied by a cost weighting coefficient. The difference between the two is the final score, used to calculate a final decision score. The simplified expression is:

[0097] Final decision score = α * Opportunity utility score - β * Execution cost score

[0098] Here, α and β are preset weight coefficients, and both are positive values. It should be noted that different weight coefficients can be set for different opportunity utility score sub-items according to actual needs. The specific setting method can be adjusted by those skilled in the art based on the actual use scenario.

[0099] After calculating the final score for all jobs in the "candidate job list", this method compares all scores and selects the job with the highest score as the unique optimal solution for the current decision period.

[0100] Based on all the information from the optimal choice, this method generates a structured operation instruction containing information such as container identification, start and end positions, and designated operating equipment, and outputs it to the corresponding equipment control interface for execution. After completing a decision, the terminal's real-time status is refreshed, and a new decision cycle begins immediately.

[0101] In addition to the above embodiments, this application also includes other implementation methods. All technical solutions formed by equivalent transformation or equivalent substitution should fall within the protection scope of the claims of this invention.

Claims

1. An automated dispatching method for loading operations at a container terminal, characterized in that, The method comprises the following steps: S1. Feasibility screening; based on the preset ship location dependence and the real-time status of the terminal yard, all currently physically executable operations are identified and collected to form a candidate operation list; The feasibility screening comprises converting the preset stowage plan into a directed acyclic graph to represent the support relationship between containers on the ship; and obtaining the coordinates, physical state and overlying container relationship information of each container to be loaded in the terminal yard in real time; The specific steps of the feasibility screening are as follows: S11. According to the directed acyclic graph, all containers whose parent nodes have all been loaded on the ship are selected from all containers to be loaded to form a logically feasible list; S12. Based on the real-time status data of the terminal yard, a physical feasibility check is performed to form a final candidate operation list, which includes: From the logically feasible list, select the containers with a physical state of being workable to be added to the candidate operation list as "loading candidate operations"; And, identify all obstacle containers in the yard whose physical state is workable but are pressed by containers to be loaded, and add them to the candidate operation list as "tumbling candidate operations"; Wherein, the condition for the physical state being workable is that there is no other container above the container, and the container is not locked by any operation process in the system; S2. Execution cost assessment; for each candidate operation in the candidate operation list, at least one preset cost dimension is quantitatively evaluated to determine the execution cost score; the cost dimension at least includes: equipment movement cost, tumbling operation cost or potential conflict cost; The potential conflict cost is proportional to the busy degree of the operation area or equipment related to the candidate container, and the quantitative indicators of the busy degree include the current instruction number of the container bay where the candidate container is located, or the current instruction load number of the responsible quay crane; S3. Opportunity utility assessment; for each candidate operation in the candidate operation list, at least one preset utility dimension is prospectively quantitatively evaluated to determine the opportunity utility score; the utility dimension at least includes: unlocking utility of unlocking subsequent key operations, synergy utility of forming double-box operations, or resource matching utility of matching with terminal resources; The unlocking utility of unlocking subsequent key operations is quantified as follows: Determine the set of containers that will be directly pressed by the container to be loaded after the execution of the current candidate operation, and accumulate the preset priority values of all containers in the set, and the accumulated total is taken as the unlocking utility score, which is used to measure the degree of advancement of the current operation to the stowage sequence; S4. Final decision and instruction generation; based on the execution cost score and the opportunity utility score, the final decision score of each candidate operation is calculated; according to the final decision score, the optimal operation is selected from the candidate operation list, and the corresponding loading operation instruction is generated.

2. The method of claim 1, wherein, The execution cost score in S2 is calculated by the following cost: The equipment movement cost is proportional to the movement distance or time of the yard crane responsible for the execution of the operation from the current real-time position to the position of the candidate container in the yard; The cost of the turning operation is evaluated according to the type of the candidate operation: For a "ship loading candidate operation", its cost is proportional to the number of other containers stacked above it that need to be temporarily moved for the extraction of the container; For a "turning candidate operation", its cost is quantified as the cost of performing the turning operation itself, and can be further evaluated for its potential impact on the future sequence of operations in the yard.

3. The method of claim 2, wherein, Different cost weights are assigned to different types of containers being turned, including normal containers, refrigerated containers, or dangerous goods containers, when calculating the cost of the turning operation.

4. The method of claim 1, wherein, The synergy effect of the double-box operation in S3 is quantified as: Detecting whether the destination location of the candidate container is adjacent or close to the destination location of another container already loaded or about to be loaded on the truck, on the ship or in the yard, and if the preset proximity condition is met, a higher synergy effect score is given.

5. The method of claim 1, wherein, The resource matching effect in S3 is quantified as: Evaluating the matching degree of the operation requirements of the candidate container and the available resources of the current terminal; when the operation is assigned to a lower load or more abundant resource quay crane, the resource matching effect score is higher.

6. The method of claim 1, wherein, The steps of the final decision and instruction generation in S4 include: S41. Calculate the final decision score for each candidate operation by a preset weighting algorithm by integrating the execution cost score and the opportunity effect score, which can be represented as: Final decision score = α * opportunity effect score - β * execution cost score Where α and β are preset weight coefficients, and both are positive values; S42. Select the operation with the highest final decision score as the optimal operation and generate the instruction.

Images