A space-time perception and multi-source conflict decoupling method for cluster scheduling of steel plant trolley buses

Abstract

The present application relates to the technical field of automation scheduling, in particular to a kind of space-time perception and multi-source conflict decoupling's steel plant crane cluster scheduling method, including obtaining business data and optimization parameter;Task execution dependency task verification, abnormal task and crane identification;Introduce multidimensional decision variable, build multi-source decoupling constraint condition, set scheduling target, build mixed integer programming model;Based on model in the set time range iteratively solves;Output crane task allocation plan, task execution plan and crane driving path, generate executable scheduling instruction, constitute crane scheduling plan, the present application can effectively handle the multi-constraint coupling and space-time conflict problem in complex industrial crane scheduling scene, effectively improve scheduling efficiency, reduce crane avoidance and collision risk, applicable to steel, port, railway freight and other complex industrial logistics scene.

Description

Technical Field

[0001] This invention is a steel plant crane cluster scheduling method that decouples spatiotemporal perception and multi-source conflict, belonging to the field of automated scheduling technology. Background Technology

[0002] Overhead cranes (bridge cranes) are core equipment in steel production logistics, undertaking critical tasks such as ladle transfer, billet loading and unloading, and equipment maintenance. With the accelerated pace of steel mill production and the increasing complexity of production lines, it has become common for multiple overhead cranes to work collaboratively in the same span. Overhead crane scheduling issues directly affect production efficiency, equipment utilization, and operational safety.

[0003] Traditional crane dispatching mainly relies on manual experience or simple rule-based scheduling, which has the following prominent problems: First, when multiple cranes work together, there are often waiting or collision risks due to path conflicts and workstation competition, affecting the continuity of operations; Second, there are complex temporal dependencies between tasks, such as lifting before moving, moving before lowering, and spatial matching requirements, such as the endpoint of the preceding task must be consistent with the starting point of the subsequent task, which makes it difficult for manual dispatching to optimize globally; Third, crane operations are subject to multiple constraints such as available time windows for tasks, maintenance plans, and safety distances, which traditional methods cannot take into account in a coordinated manner.

[0004] Existing automated scheduling methods mostly employ mathematical programming or heuristic algorithms, but they still face the following technical bottlenecks: First, the models do not characterize spatiotemporal resources precisely enough, making it difficult to accurately describe the dynamic evolution of workstation occupancy status as tasks are executed; second, they lack the ability to decouple multi-source conflicts, such as task conflicts, path conflicts, and resource conflicts, and often can only handle single-type conflicts; third, they lack effective modeling of actual process requirements such as uninterrupted tasks and crane position locking, resulting in poor executability of scheduling schemes.

[0005] Therefore, this invention proposes an intelligent scheduling method for train clusters that integrates spatiotemporal perception and multi-source conflict decoupling, thereby achieving global optimization and dynamic coordination of multi-train operations and improving the intelligence level of steel plant logistics systems. Summary of the Invention

[0006] To address the problems in existing technologies, this invention provides a steel plant crane cluster scheduling method that decouples spatiotemporal perception from multi-source conflicts.

[0007] The technical solution adopted by this invention to solve its technical problem is: a steel plant crane cluster scheduling method for decoupling spatiotemporal perception and multi-source conflict, comprising: Step S1: Obtain business data and optimization parameters; Step S2: Perform task dependency verification, abnormal task and vehicle recognition; Step S3: Introduce multidimensional decision variables, construct multi-source decoupling constraints, set scheduling objectives, and construct a mixed integer programming model; Step S4: Iteratively solve the problem based on the model within a set time range; Step S5: Output the driving task allocation plan, task execution plan and driving route, generate executable scheduling instructions, and form a driving scheduling plan.

[0008] Furthermore, step S1 involves acquiring data from the vehicle control system and the production system.

[0009] Furthermore, the dependency task verification includes: Spatial consistency check: If a task has dependent tasks, check whether the starting coordinates are equal to the ending coordinates of the dependent tasks. If they are not equal, issue a warning and make corrections. Dependency existence check: If the task must be executed continuously, it must be associated with a dependent task; otherwise, an exception is thrown. The abnormal task and vehicle recognition include abnormal task recognition and abnormal vehicle recognition; The abnormal task identification includes: Determine if the sum of the earliest start time and the execution time of a task is less than the maximum planned duration. If it is greater, identify the task as an abnormal task and remove it. Determine whether the starting and ending workstation positions of the task are outside the movable range of the crane. If so, identify it as an abnormal task and remove it. Abnormal vehicle identification includes checking the vehicle's fault and maintenance status within the maximum planned duration, identifying unavailable fault markers, processing maintenance data into the algorithm's input data format, and applying maintenance constraints to fix the position.

[0010] Furthermore, the introduced multidimensional decision variables include workstation occupancy status variables, task start and completion variables, crane displacement variables, task delay status variables, and related constants; The multi-source decoupling constraints include task timing, vehicle movement, spatiotemporal safety, task attributes, vehicle status, complex task groups, workstation binding, and workstation execution occupancy constraints. The scheduling objectives include minimizing the total task delay penalty considering priority and minimizing the travel distance of the vehicle cluster.

[0011] Furthermore, each of the tasks has a specified start and end point and is completed within a specified time window.

[0012] Furthermore, the tasks are performed in order of priority, with high-priority tasks to be executed first, and tasks currently being executed at the initial moment to be completed first.

[0013] Furthermore, the task operation process must satisfy constraints on its own movable range, the uniqueness of the task workstation, the spatiotemporal safety, and the dependency relationship between tasks.

[0014] Furthermore, the multi-source decoupling constraints include: The task uniqueness constraint is expressed mathematically as follows: The mathematical expression for the task priority constraint during execution is: Depending on the spatiotemporal constraints of the task, the mathematical expression is: The mathematical expression for the vehicle displacement constraint is: The mathematical expression for the vehicle speed limit constraint is: The task execution time window constraint, mathematically expressed as: The task execution state constraint, expressed mathematically as follows: The job position matching constraint, expressed mathematically as follows: The job position locking constraint, mathematically expressed as: The spatiotemporal safety collision avoidance constraint is expressed mathematically as follows: The initial position constraint for vehicle movement is expressed mathematically as follows: The mathematical expression for the vehicle's movable range constraint is: The mathematical expression for the task's continuous execution constraint is: : The maintenance condition constraint, mathematically expressed as: The workstation occupancy constraint, expressed mathematically as follows: .

[0015] Furthermore, the scheduling objective includes: The mathematical expression for minimizing the total task delay penalty considering priority is: The mathematical expression for minimizing the travel distance of the vehicle cluster is: The overall objective, expressed mathematically, is: .

[0016] The beneficial effects of this invention are: 1. By using spatiotemporal perception and multi-source conflict decoupling mechanisms, multi-vehicle collaborative optimization is achieved, significantly reducing task waiting and vehicle avoidance times, shortening task completion time, improving the feasibility of the solution, and improving scheduling efficiency; 2. Strictly adhere to safe driving distances, unique workstation occupancy, and locked maintenance locations to effectively avoid collisions and resource conflicts, ensuring the continuity and safety of the production process and guaranteeing operational safety; 3. It reduces human intervention, improves the level of automated scheduling, makes intelligent decisions quickly, and can be extended to similar complex logistics scenarios such as steel, ports, railway freight and other complex industrial logistics scenarios, with a wide range of applications. Attached Figure Description

[0017] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the method flow for a steel plant crane cluster scheduling method that decouples spatiotemporal perception and multi-source conflict according to the present invention. Detailed Implementation

[0018] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0019] Please see Figure 1 This embodiment provides a steel plant crane cluster scheduling method that decouples spatiotemporal awareness from multi-source conflicts, including: Step S1: Obtain business data and optimization parameters; Step S1: Obtain data from the vehicle control system and production system, and optimize parameters to support manual input.

[0020] Step S2: Verify task dependencies for task execution and identify abnormal task driving conditions; Dependency task verification includes: Step 1: Spatial consistency check. If the task has dependent tasks, check whether the starting coordinates are equal to the ending coordinates of the dependent tasks. If they are not equal, issue a warning and correct it. Step 2: Dependency existence check. If the task must be executed continuously, it must be associated with a dependent task; otherwise, throw an exception. Preferably, the tasks that must be executed continuously are classified by dependency level, including tasks with strong continuous hard dependencies and tasks with weak continuous soft dependencies. Different dependency verification and task release strategies are matched for continuous tasks of different levels. Among them, hard-dependent tasks are continuous operation tasks corresponding to the core processes of the steel plant, such as continuous casting ladle transfer and refining molten steel continuous hoisting tasks. There is no fault tolerance space for the continuity of such tasks. They must be associated with complete and executed preceding dependent tasks. If no valid preceding dependent task is matched, an exception will be thrown directly and the scheduling of the current continuous task will be blocked. Soft-dependent tasks are continuous operation tasks that correspond to auxiliary production in steel plants, such as continuous replenishment of raw materials in the storage area, continuous return of crane equipment, and continuous removal of waste materials. These tasks do not affect the production continuity of the main steelmaking process and have the characteristics of flexible and adjustable timing. If a soft-dependent task does not match a preceding dependent task, it will not throw an exception or intercept task scheduling, and will allow the task to start directly and adapt to the workshop conditions for execution.

[0021] Soft-dependent tasks always retain the continuous operation attribute. Under normal working conditions, when there are legitimate preceding dependent tasks, soft-dependent tasks are executed sequentially following the preceding tasks to maintain the batch continuous operation rhythm. Only in abnormal working conditions without dependencies, such as initial startup, task queue clearing, or abnormal termination of preceding tasks, is the flexible release logic executed. After the task starts, the continuous connection of subsequent tasks can be automatically rebuilt to restore the continuous operation mode.

[0022] Abnormal tasks and vehicle recognition include: Step 1: Abnormal task identification. Determine if the earliest start time plus the execution time of the task is less than the maximum planned duration. If it is greater, the task is identified as abnormal and removed. Determine if the start and end workstation positions of the task exceed the movable range of the crane. If they do, the task is identified as abnormal and removed. Step 2: Abnormal driving identification. Check the vehicle's fault and maintenance status within the maximum planned time. Fault markers are unavailable. Maintenance data is processed into the algorithm input data format and maintenance constraints are applied to fix the position.

[0023] Step S3: Introduce multidimensional decision variables, construct multi-source decoupling constraints, set scheduling objectives, and construct a mixed integer programming model; Multidimensional decision variables are introduced, including workstation occupancy status variables, task start and completion variables, crane displacement variables, task delay status variables, and related constants. The multi-source decoupling constraints include constraints on task timing, vehicle movement, spatiotemporal safety, task attributes, vehicle status, complex task groups, workstation binding, and workstation execution occupancy. Setting scheduling objectives includes minimizing the total task delay penalty considering priority and minimizing the travel distance of the vehicle cluster.

[0024] Step S4: Iteratively solve the problem based on the model within a set time range; Step 5: Output the driving task allocation plan, task execution plan and driving route, generate executable scheduling instructions, and form a driving scheduling plan.

[0025] The steel plant crane cluster scheduling problem addressed by this invention is as follows: Within a horizontal span of a steel mill workshop, several tasks are distributed for execution. Each task has a designated start and end point, as well as a strict time window: the earliest start time and the latest end time. Tasks must not be executed earlier than the earliest start time and must be completed no later than the latest end time, if possible. The urgency of a task is indicated by its priority; higher priority tasks should be executed first, and tasks currently in progress must be completed first.

[0026] Multiple overhead cranes are deployed within the area, initially randomly distributed. Crane movement is limited by maximum speed and can only operate within a pre-defined reachable range. To ensure safe production, any two overhead cranes must maintain a distance of at least a safe clearance during movement. During task execution, the operation cannot be interrupted, paused, or interspersed with other tasks to ensure the continuity of the lifting process. Furthermore, during pre-defined maintenance periods, the relevant overhead cranes must not move. In future scheduling periods, a collaborative operation plan for the crane cluster must be developed to meet all the above constraints and achieve the following optimization objectives: high-priority tasks should be executed as early as possible; the overall delay of all tasks should be minimized; and unnecessary movement of overhead cranes due to obstacle avoidance should be minimized, thereby optimizing the total travel distance. The scheduling plan must be generated within seconds to meet the needs of real-time dynamic scheduling.

[0027] Based on the above issues, this business problem is viewed as follows: Each task is defined as a directed path segment, from the starting workstation to the ending workstation, representing a complete operation that the train needs to complete. The train scheduling plan can be viewed as starting from the initial positions of each train and finding a set of spatiotemporal paths covering as many tasks as possible. The paths must sequentially pass through the starting and ending points of the assigned tasks, ultimately reaching a certain ending position. These paths must satisfy task priority order, with higher-priority tasks scheduled first. The uniqueness of workstation occupancy, safe train distance, task time windows, and task dependencies are used as constraints on the spatiotemporal paths. Collaborative optimization scheduling of the train cluster is achieved through multi-source conflict decoupling.

[0028] The mixed-integer linear model constructed in this invention, wherein: Multidimensional decision variables include: Indicates driving Is it in Start executing the task at any time ; Indicates driving Is it in End the task at the right time ; Indicates the position of vehicle j at time t; Indicates that vehicle j is in the time period. The distance traveled within the vehicle's path is used to quantify the length of the driving route. Indicates task The actual delay time; Indicates task Has the execution been completed? Indicates task Whether to start execution; Indicates workstation exist Is the time slot occupied? Relevant constants include: Indicates task The starting workstation location; Indicates task The final workstation location; Indicates task The latest end time; Indicates task The earliest end time; Indicates task Execution time at the starting workstation; Indicates task Execution time at the final workstation; Indicates task Execution priority; Indicates driving The initial position; Indicates driving Maximum driving speed; Indicates driving The task being performed at the initial moment; Indicates driving Execute the task Required time unit; Indicates driving The start time of the maintenance; Indicates driving The maintenance completion time; Indicates driving The leftmost position of the span that can be moved; Indicates driving The rightmost position that can be moved to; Indicates driving The location for maintenance; Indicates task The dependent tasks; This represents the set of tasks to be scheduled. Represents the set of all tasks; This represents a very large positive number; Indicates the planned scheduling duration; Indicates the safe distance between lines; The constraint expressions involved in multi-source decoupling constraints are as follows: Task uniqueness constraint: It is used to ensure that each task can only be started by one vehicle at one time, and that each vehicle can handle at most one task at the same time, preventing tasks from being repeatedly assigned or vehicle conflicts.

[0029] Task priority constraints during execution: It is used to ensure that the tasks being performed at the beginning of the journey are executed first and are not preempted or delayed by subsequent tasks.

[0030] Depends on the spatiotemporal constraints of the task: It is used to ensure that the task and its dependent tasks must be executed by the same machine, and that the task starts later than the dependent task finishes.

[0031] Vehicle displacement constraints: It is used to calculate the driving speed of the vehicle. The distance traveled within the time frame is equal to the positional difference between adjacent time points.

[0032] Vehicle speed limit constraints: It is used to limit the speed of vehicle movement to ensure that the displacement difference does not exceed the speed limit within a single unit of time.

[0033] Task execution time window constraints: It is used to ensure that the start time of a task is not earlier than the earliest allowed start time, and to record the offset of the task's completion time from the given latest completion time.

[0034] Task execution state constraints: It is used to record the start / end execution status of tasks, ensuring that each task has a unique start and end point.

[0035] Job location matching constraints: It is used to restrict the vehicle to a fixed position at the start / end point after performing a task, ensuring that the task is performed correctly in space.

[0036] Job position locking constraints: Its combination of constraint-based job position matching ensures that the crane remains in a fixed position during the execution of tasks at the starting / ending workstations, and does not allow displacement during the operation, thus ensuring operational safety and process requirements.

[0037] Spatiotemporal safety collision avoidance constraints: It is used to ensure that all vehicles maintain a safe distance at all times, thus guaranteeing the physical safety of multiple vehicles operating in coordination.

[0038] Initial position constraints for vehicle movement: It sets the spatial coordinates of the scheduling start time for each train to ensure that the physical state of the model can be deduced at the initial moment.

[0039] Vehicle movement range constraints: It is used to ensure that the position of each vehicle does not deviate from the vehicle's movable range at any given time; Task continuation constraints: : It is used to ensure that once a task is started by a crane, it must be carried out continuously until completion, without interruption, suspension or interleaving of other tasks, thus ensuring the continuous process of hoisting.

[0040] Maintenance status constraints: It restricts the operation of vehicles within the maintenance time window; vehicles must not be moved and must remain stationary at the designated maintenance point to ensure the safety and planning of maintenance work.

[0041] Workstation occupancy constraints: The system restricts workstations from being occupied and released simultaneously, ensuring the exclusivity of each workstation.

[0042] Regarding the overall scheduling objectives: Minimize the total task delay penalty considering priority: Minimize the travel distance of the vehicle cluster: Overall Objective: This invention can effectively handle the multi-constraint coupling and spatiotemporal conflict problems in complex industrial train scheduling scenarios, effectively improve scheduling efficiency, reduce train avoidance and collision risks, and is applicable to complex industrial logistics scenarios such as steel, ports, and railway freight.

[0043] This embodiment also provides a method for scheduling steel plant crane clusters by decoupling spatiotemporal awareness and multi-source conflict, including the following steps: Step 1: Data acquisition, including business data and optimization parameters; Driving and task information are obtained from the vehicle control system and production system. Specific driving-related input data are shown in the table below: Table 1 Driving Information The task-related input data is shown in the table below: Table 2 Task Information The parameter configuration input data is shown in the table below: Table 3 Parameter Configuration Step 2: Task preprocessing, including dependency task verification and abnormal task / vehicle identification; First, we performed a dependency task check. There were no task dependencies, so the check passed.

[0044] Next, abnormal task / vehicle recognition is performed.

[0045] First, check if the earliest start time plus execution time of the task is less than the maximum planned duration. The latest durations of the tasks are 4, 5, 5.88, 5.88, 13.31, 23.5, 16.23, 20.6, 10.24, 24.6, 26, 14.11, 13.51, 5, and 6, all of which are less than 45, so the check passes.

[0046] Step 3: Construction of the mixed-integer programming model, including the introduction of multi-dimensional decision variables, the construction of multi-source decoupling constraints, and the setting of scheduling objectives; The specific objective function for the total task delay penalty is as follows: The specific objective function for the vehicle movement penalty involved is shown below: The specific constraint expressions involved are shown below: Task uniqueness constraint: Task priority constraints during execution: The SGK03 vehicle must complete the currently executing task 1 before executing other tasks; Depends on the spatiotemporal constraints of the task: Constraint generation is not performed on tasks without dependencies.

[0047] Vehicle displacement constraints: Vehicle speed limit constraints: Task execution time window constraints: The actual start time of tasks 2 and 3 must not be earlier than task 2, and the actual start time of tasks 4 and 5 must not be earlier than task 3; calculate the delay time for each task. ; Task execution state constraints: Job location matching constraints: The vehicle is restricted to its original position at a fixed time after completing the task at the starting / ending workstation to ensure spatial accuracy during task execution. (Task 1 is used as an example.) Job position locking constraints: By combining the work position matching constraints, the crane is ensured to remain in a fixed position during the execution of the task at the starting / ending work station, and displacement is not allowed during the operation, so as to ensure the safety of the operation and process requirements. Spatiotemporal safety collision avoidance constraints: Ensure that the distance between all vehicles at any given time does not exceed a safe distance of 5 meters; Initial position constraints for vehicle movement: Vehicle movement range constraints: Task continuation constraints: Ensure that once a task is started by a crane, it must be carried out continuously until completion, without interruption, pause, or interleaving with other tasks, to guarantee the continuous process requirements of hoisting. Maintenance status constraints: No constraints are generated when a vehicle is not inspected and repaired.

[0048] Workstation occupancy constraints: The system restricts workstations from being occupied and released simultaneously, ensuring the exclusivity of each workstation.

[0049] Step 4: Model solving. Set a time limit of 30 seconds and use 4 threads to call CPLEX for solving. The results are shown in Table 4. Table 4 Task Allocation Results Step 5: Output Results. The output results are divided into two categories: one is the task allocation plan. The output is shown in Table 1, including the task number, execution start time, and execution end time. In addition, a driving route table will also be output; for example, Table 5 shows the route for the past ten days.

[0050] Table 5 Driving Route Table This embodiment offers advantages such as effective conflict decoupling, priority guarantee, significant path optimization, and fast and reliable solution. Specifically, under the premise of meeting safety distance constraints, multi-vehicle collaborative operations are collision-free, successfully avoiding spatiotemporal conflicts between tasks and ensuring production safety. High-priority tasks are prioritized for execution, and their completion time is strictly controlled within the time window, reflecting a priority-oriented scheduling logic. The total travel distance and number of avoidance maneuvers are reduced, improving equipment utilization and operational efficiency. Feasible scheduling solutions are obtained within seconds, meeting the requirements of real-time dynamic scheduling.

[0051] Although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for scheduling steel plant crane clusters based on spatiotemporal awareness and decoupling of multi-source conflicts, characterized in that: include: S1. Obtain business data and optimization parameters; S2. Task execution depends on task verification, abnormal tasks, and vehicle recognition; S3. Introduce multidimensional decision variables, construct multi-source decoupling constraints, set scheduling objectives, and construct a mixed integer programming model; S4. Iteratively solve the model within a set time range; S5 outputs the driving task allocation plan, task execution plan, and driving route, generating executable scheduling instructions to form a driving scheduling plan.

2. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 1, characterized in that: S1 obtains data from the vehicle control system and the production system.

3. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 2, characterized in that: The dependency task verification includes: Spatial consistency check: If a task has dependent tasks, check whether the starting coordinates are equal to the ending coordinates of the dependent tasks. If they are not equal, issue a warning and make corrections. Dependency existence check: If the task must be executed continuously, it must be associated with a dependent task; otherwise, an exception is thrown. The abnormal task and vehicle recognition include abnormal task recognition and abnormal vehicle recognition; The abnormal task identification includes: Determine if the sum of the earliest start time and the execution time of a task is less than the maximum planned duration. If it is greater, identify the task as an abnormal task and remove it. Determine whether the starting and ending workstation positions of the task are outside the movable range of the crane. If so, identify it as an abnormal task and remove it. Abnormal vehicle identification includes checking the vehicle's fault and maintenance status within the maximum planned duration, identifying unavailable fault markers, processing maintenance data into the algorithm's input data format, and applying maintenance constraints to fix the position.

4. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 1, characterized in that: The introduced multidimensional decision variables include workstation occupancy status variables, task start and completion variables, vehicle displacement variables, task delay status variables, and related constants; The multi-source decoupling constraints include task timing, vehicle movement, spatiotemporal safety, task attributes, vehicle status, complex task groups, workstation binding, and workstation execution occupancy constraints. The scheduling objectives include minimizing the total task delay penalty considering priority and minimizing the travel distance of the vehicle cluster.

5. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 1, characterized in that: Each task has a specified start and end point and must be completed within a specified time window.

6. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 5, characterized in that: The tasks are performed in order of priority, with high-priority tasks to be executed first, and tasks that are currently being executed at the initial moment to be completed first.

7. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 1, characterized in that: The task operation process must meet the following constraints: its own movable range, the uniqueness of the task workstation, the spatiotemporal safety, and the dependency relationship between tasks.

8. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 4, characterized in that: The multi-source decoupling constraints include: The task uniqueness constraint is expressed mathematically as follows: The mathematical expression for the task priority constraint during execution is: Depending on the spatiotemporal constraints of the task, the mathematical expression is: The mathematical expression for the vehicle displacement constraint is: The mathematical expression for the vehicle speed limit constraint is: The task execution time window constraint, mathematically expressed as: The task execution state constraint, expressed mathematically as follows: The job position matching constraint, mathematically expressed as: The job position locking constraint, mathematically expressed as: The spatiotemporal safety collision avoidance constraint is expressed mathematically as follows: The initial position constraint for vehicle movement is expressed mathematically as follows: The mathematical expression for the vehicle's movable range constraint is: The mathematical expression for the task's continuous execution constraint is: ： The maintenance condition constraint, mathematically expressed as: The workstation occupancy constraint, expressed mathematically as follows: 。 9. The steel plant crane cluster scheduling method based on spatiotemporal perception and multi-source conflict decoupling as described in claim 8, characterized in that: The scheduling objectives include: The mathematical expression for minimizing the total task delay penalty considering priority is: The mathematical expression for minimizing the travel distance of the vehicle cluster is: The overall objective, expressed mathematically, is: 。

Images