Intelligent control method for dynamic tasks of static meteorological satellite

Abstract

The invention relates to the technical field of intelligent control, in particular to a static meteorological satellite dynamic task intelligent control method, which enables a static meteorological satellite ground system to realize automatic execution and quick response of a dynamic task through an intelligent control algorithm and flow when facing the dynamic task. And the requirements in the aspects of emergency response requirements, meteorological service guarantee, satellite safety management and the like are effectively met. According to the intelligent control strategy, the intelligent level and the operation efficiency of the system are improved, and more accurate and efficient services are provided for meteorological observation and satellite operation; meanwhile, diversified meteorological service requirements can be better met, the efficiency and reliability of meteorological observation and disaster early warning are improved, and the risk of manual operation is reduced.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to an intelligent control method for dynamic missions of geostationary meteorological satellites. Background Technology

[0002] With the significant improvement in remote sensing satellite payloads' remote sensing observation capabilities and instrument performance, higher demands are being placed on the precision and intelligence of ground operation control. Furthermore, the increasing frequency of meteorological disasters and weather events is placing higher demands on the emergency response capabilities of satellite observation, including timeliness, intelligence, and reliability. Moreover, the continuous enrichment of satellite observation resources and the diversification of remote sensing applications undoubtedly increase the complexity and challenges of mission management and control for satellite ground systems.

[0003] Satellite ground systems are crucial for satellite data reception, preprocessing, and system integration. Currently, satellites in different fields within China are gradually developing and constructing adaptive ground systems based on their specific characteristics and functions. The HY-2 ocean satellite ground system is responsible for the operational management of the HY-2 ocean exploration satellite, the generation of ocean satellite remote sensing products, and its application demonstration services. The Gaofen satellite ground system is a vital support system connecting Gaofen satellites and applications, used for satellite data entry, production, archiving, simulation, distribution, calibration of Level 1-2 standard products, and product sharing services. The Fengyun-2 satellite has a single payload and a simple ground control design. In summary, remote sensing satellite ground systems focus on the management and control of remote sensing satellite missions, ensuring data reception and transmission, and the production, storage, and distribution of standard products.

[0004] After conducting thorough research on different satellite ground systems, it was found that there is a lack of research on how to intelligently execute dynamic missions in existing satellite ground systems, especially a lack of solutions for intelligent control of dynamic missions in geostationary meteorological satellite ground systems.

[0005] Based on the above reasons, this invention designs an intelligent control method for dynamic missions of geostationary meteorological satellites. Combining the characteristics and functions of Chinese geostationary meteorological satellites, it proposes an intelligent scheduling and control algorithm for dynamic missions adapted to geostationary meteorological satellite ground systems. This enables the mission-driven ground system to operate intelligently when facing dynamic missions and flexibly respond to various complex situations. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent control method for dynamic missions of geostationary meteorological satellites. Combining the characteristics and functions of geostationary meteorological satellites themselves, this invention proposes an intelligent scheduling and control algorithm for dynamic missions adapted to geostationary meteorological satellite ground systems. This enables the mission-driven ground system to operate intelligently when faced with dynamic missions and flexibly respond to various complex situations.

[0007] To achieve the above objectives, the present invention provides an intelligent control method for dynamic missions of geostationary meteorological satellites, comprising the following steps:

[0008] S1 generates a dynamic task schedule by modifying the task schedule template. Driven by the dynamic task schedule, the ground system and its scheduling control are triggered through the intelligent control platform.

[0009] S2. Based on the characteristics of dynamic services, a differentiated scheduling strategy is adopted. For dynamic services with high timeliness requirements, integrated scheduling is adopted, while for dynamic services with low real-time requirements, hierarchical scheduling is adopted.

[0010] S3, through in-depth analysis of ground mission control and command uploading execution time logic, designs scheduling methods and algorithms to achieve optimal system performance;

[0011] S4, based on the dynamic task application time and execution time requirements, the intelligent control platform selects different control processes, specifically:

[0012] Control flow A: t q ∈[T0+24h,T0+48h] and t s <T0;

[0013] Control Flow B: t q ∈[T0+24h,T0+48h] and T0 <t s <T1-Δt_ctrl;

[0014] Control flow C: t q ∈[T0+24h,T0+48h] and T1-Δt_ctrl <t s , or t s With t q On the same day;

[0015] Among them, t s It is the time for applying for the observation mission, t q T0 is the task execution time. The system switches to dynamic task working mode. T1 is the time when the task schedule is generated and issued for the next day. Δt_ctrl is the minimum control cycle of the dynamic task.

[0016] S5, based on the intelligent control process of S4, controls the ground system according to the characteristics of dynamic business:

[0017] For hierarchical scheduling: the ground system distributes task schedules to various technical systems through message queues and issues commands to other systems using standardized scheduling orders. The commands include two types of scheduling orders: task schedule arrival or task schedule update. After receiving the task schedule, the technical system needs to send a confirmation reply to the task management and control system and return communication confirmation information after receiving the scheduling order.

[0018] For integrated scheduling: the ground system maximizes the task priority benefit f1 and minimizes the task execution time deviation f2 and the computational resource usage cost f3. It adopts a multi-objective evolutionary algorithm combined with constraints to perform multi-objective optimization in order to achieve the optimal global scheduling of the system. A penalty function is introduced to punish individuals that violate the constraints, which is used to guide the multi-objective evolutionary algorithm to search for feasible solutions.

[0019] S6, Calculation of the objective function:

[0020] Let X = [x1, x2, ..., xn] N ], where x i ∈{0,1} indicates whether task i is scheduled, with 1 for scheduled and 0 for unscheduled;

[0021] Let S = [S1, S2, ..., S2] N [] indicates the start time of each task;

[0022] Let R = [r1, r2, ..., r] N ] represents the resource combination allocated to each task;

[0023] The objective function is then:

[0024] Formula 1:

[0025] Formula 2:

[0026] Formula 3:

[0027] Among them, w i The weight of task i to be executed, t qi Let i be the expected execution time;

[0028] S7, the multi-objective evolutionary algorithm is as follows:

[0029] S7-1, based on a multi-objective optimization model, initializes the population and calculates the objective function value. Here, the population is the set of candidate solutions.

[0030] S7-2, stratification is performed through non-dominated sorting to generate uniformly distributed reference points and establish the association between individuals and reference points;

[0031] S7-3 selects the next generation of the population through an elite strategy and performs crossover and mutation operations on various populations;

[0032] S7-4, repeat S7-1 to S7-3 until the termination condition is met.

[0033] The mission schedule templates include imager mission schedule templates, detector mission schedule templates, rapid imager mission schedule templates, and satellite orbit determination and management mission schedule templates, etc.

[0034] The task schedule template is defined by task and uses the Extensible Markup Language (XML) file format specification. The format specification includes task sequence number, task name in English, task number, task start time, task end time, and task parameters.

[0035] In S4, T0 is daily UTC 04:00; T1 is daily UTC 22:00; Δt_ctrl is 10 minutes.

[0036] The penalty function in the integrated scheduling of S5 is as follows:

[0037] S5-1, Constraints on the time window:

[0038] Start time S of task i i It should be within its permitted time window, that is:

[0039] Formula 4: t qi ≤S i ≤t qi +Δt window ; where Δt window The time window length for task i. If task i is scheduled but exceeds the time window, a penalty term is added:

[0040] Formula 5:

[0041] S5-2, Constraints on Resources:

[0042] The total combination of resources allocated to tasks cannot exceed the system's total resource limit R. max ,Right now:

[0043] Formula Six: If the total resources exceed the limit, a penalty will be applied:

[0044] Formula 7:

[0045] S5-3, Dependencies on Tasks:

[0046] If task j depends on task k, then the end time of task k must be earlier than the start time of task j.

[0047] Formula 8: S i ≥S k +Δt k ; where Δt k For the execution time of task k, if the dependency is not satisfied, a penalty term is added:

[0048] Formula 9: P dependency =∑ 所有依赖关系 max(0,S j -(S k +Δt k ));

[0049] S5-4, Definition of the comprehensive penalty function:

[0050] After weighting the penalty terms from S5-1 to S5-3, and adding them to the objective function, we obtain the comprehensive penalty function:

[0051] Formula 10: P = αP time +βP resource +γP dependency ; where α, β, and γ are weighting coefficients, which are dynamically adjusted according to the load of the ground system, the type of dynamic mission, and the number of iterations.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] This invention enables geostationary meteorological satellite ground systems to automatically execute and rapidly respond to dynamic tasks through intelligent control algorithms and processes, effectively meeting the requirements of emergency response, meteorological service support, and satellite safety management. Its intelligent control strategy not only improves the system's intelligence level and operational efficiency, providing more accurate and efficient services for meteorological observation and satellite operation, but also better meets diverse meteorological service needs, enhancing efficiency and reliability in meteorological observation and disaster early warning, and reducing the risk of human error. Attached Figure Description

[0054] Figure 1 This is a diagram of the intelligent autonomous control framework for a geostationary meteorological satellite ground system mission disclosed in an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of the intelligent control process matching for geostationary meteorological satellites disclosed in an embodiment of the present invention;

[0056] Figure 3 This is a flowchart of the geostationary meteorological satellite payload observation mode switching management disclosed in an embodiment of the present invention;

[0057] Figure 4This is a schematic diagram of the intelligent control process A of the geostationary meteorological satellite ground system disclosed in an embodiment of the present invention;

[0058] Figure 5 This is a schematic diagram of the intelligent control process B of the geostationary meteorological satellite ground system disclosed in an embodiment of the present invention;

[0059] Figure 6 This is a schematic diagram of the intelligent control process C of the geostationary meteorological satellite ground system disclosed in an embodiment of the present invention;

[0060] Figure 7 This is a flowchart illustrating the implementation of the NSGA-III algorithm for a geostationary meteorological satellite ground system as disclosed in an embodiment of the present invention.

[0061] Figure 8 This is a flowchart of the control process for switching operational observation modes of geostationary meteorological satellites, as disclosed in an embodiment of the present invention. Detailed Implementation

[0062] The present invention will now be further described with reference to the accompanying drawings.

[0063] See Figures 1-8 This invention provides an intelligent control method for dynamic missions of geostationary meteorological satellites:

[0064] When dynamic tasks need to be executed, a dynamic task schedule is generated by modifying the task schedule template. Driven by the dynamic task schedule, the ground system and its scheduling control are triggered through the intelligent control platform.

[0065] Differentiated scheduling strategies are adopted based on business characteristics. Integrated scheduling is used for business systems and jobs with high timeliness requirements, while hierarchical scheduling control is used for those with lower real-time requirements. Integrated scheduling and centralized management facilitate system performance optimization and unified management throughout the entire process. Hierarchical scheduling ensures that each level does not interfere with the others and has the capability for reprocessing scheduling control.

[0066] Due to the diversity of tasks, different task schedule templates are introduced into the geostationary meteorological operational system, including imager task schedule templates, detector task schedule templates, rapid imager task schedule templates, and satellite orbit determination and management task schedule templates.

[0067] These task schedule templates are defined according to a unified standardized file format, facilitating the reading, parsing, maintenance, management, and application of task schedules. The task schedule templates are defined on a task-by-task basis using the Extensible Markup Language (XML) file format specification, and include attributes such as task sequence number, task name, task number, task start time, task end time, and task parameters.

[0068] For example, the imager mission schedule template format is as follows:

[0069]

[0070]

[0071] In the dynamic task working mode, changes occur in each link of task planning, plan formulation, and instruction generation compared to regular tasks. Therefore, it is necessary to deeply analyze the ground task control and the time logic of instruction uploading and execution, and design a reasonable scheduling method and algorithm.

[0072] For the convenience of description, several time variables related to task planning and scheduling are defined below.

[0073] ts: The application time for the observation task.

[0074] tq: The task execution time when the system switches to the dynamic task working mode.

[0075] T0: The generation and distribution time of the next-day task schedule. The next-day task schedule is generated at UTC 04:00 every day and distributed to each subsystem of the ground system.

[0076] T1: The start time of uploading the next-day instructions. The next-day instructions start to be uploaded at UTC 22:00 every day.

[0077] Δt_ctrl: The minimum control period for dynamic tasks. The current system setting is 10 minutes.

[0078] According to the task application time and execution time requirements, the intelligent control platform selects different task scheduling methods and algorithms, such as Figure 2 As shown in the figure, the intelligent control processes A, B, and C in the figure have different control processes according to different dynamic task control time periods. Specifically:

[0079] (1) If t q ∈[T0 + 24h, T0 + 48h] and t s < T0; the dynamic task intelligent control executes control process A.

[0080] (2) If t q ∈[T0 + 24h, T0 + 48h] and T0 < t s < T1 - Δt_ctrl; the dynamic task intelligent control executes control process B.

[0081] (3) If t q ∈[T0 + 24h, T0 + 48h] and T1 - Δt_ctrl < t s , or t s and t q are on the same day, the dynamic task intelligent control executes control process C.

[0082] 3. Dispatch and control

[0083] When a dynamic mission is initiated, the ground system is controlled based on the intelligent control process, driven by the mission's timeline.

[0084] For hierarchical scheduling, the ground system distributes task schedules to various technical systems such as positioning and calibration through message queues, and issues commands to other systems using standardized scheduling orders, including two types of scheduling orders: task schedule arrival or task schedule update.

[0085] Upon receiving the task schedule, each technical system must send a confirmation reply to the task management and control system, and return a communication confirmation message upon receiving the dispatch order. This dispatch mechanism achieves inter-system collaboration through standardized command formats, where the format definition of the dispatch order is shown in Table 1, and an example of the dispatch order content definition is shown in Table 2.

[0086] Table 1. Dispatch Order Data Field Format Definition

[0087]

[0088] Table 2 Definition of Dispatch Order Content

[0089]

[0090] For integrated scheduling, the system achieves optimal global scheduling by maximizing task priority gains f1 and minimizing task execution time deviations f2 and computational resource usage costs f3.

[0091] Based on this, the system employs the multi-objective evolutionary algorithm (NSGA-III) for multi-objective optimization. The system's task scheduling is subject to certain constraints, such as time windows and resource limitations. To ensure that the solutions generated by the NSGA-III algorithm conform to these constraints, a penalty function is introduced to penalize individuals that violate the constraints, thereby guiding the algorithm towards a feasible solution.

[0092] The penalty function is designed as follows:

[0093] 1) Time window constraint

[0094] The start time Si of task i must be within its allowed time window, that is:

[0095] t qi ≤S i ≤t qi +Δtwindow;

[0096] Where, Δt window The time window length for task i. If task i is scheduled but exceeds the time window, a penalty term is added:

[0097]

[0098] 2) Resource constraints

[0099] The total combination of resources allocated to tasks cannot exceed the system's total resource limit Rmax, that is:

[0100]

[0101] If the total resources exceed the limit, a penalty will be applied:

[0102]

[0103] 3) Task Dependencies

[0104] If some tasks depend on the completion of other tasks, then it is necessary to ensure that the dependent tasks are executed first. For example, if task j depends on task k, then the end time of task k must be earlier than the start time of task j.

[0105] S i ≥S k +Δt k ;

[0106] Where, Δt k Let this be the execution time of task k. If the dependency is not satisfied, a penalty term is added:

[0107] P dependency =∑ 所有依赖关系 max(0,S j -(S k +Δt k ));

[0108] 4) Comprehensive penalty function

[0109] After weighting the above penalty terms, add them to the objective function. A comprehensive penalty function can be defined as follows:

[0110] P = αP time +βP resource +γP dependency ;

[0111] Wherein, α, β, and γ are weighting coefficients, and the penalty coefficients are dynamically adjusted according to the system load, task type, and number of iterations.

[0112] Let X = [x1, x2, ..., xN] where x i Let {0, 1} represent whether task i is scheduled (1 for scheduled, 0 for unscheduled); S = [S1, S2, ..., SN] represent the start time of each task; R = [r1, r2, ..., rN] represent the resource combination allocated to each task. Then the objective function is:

[0113]

[0114] Where Wi is the weight of task i and tqi is the expected execution time of task i.

[0115] Based on the above definitions, the NSGA-III algorithm, grounded in a multi-objective optimization model, first initializes the population and calculates the objective function value. It then performs stratification through non-dominated sorting, generates uniformly distributed reference points, establishes associations between individuals and these reference points, selects the next generation of the population using an elitist strategy, and finally performs crossover and mutation operations on the population. This process is repeated until a termination condition is met. Its implementation flow is as follows: Figure 1 As shown.

[0116] Based on actual business needs, the optimal scheduling scheme can be selected from the Pareto frontier: for example, when considering satellite service telemetry and control, the scheduling scheme that simultaneously satisfies time window constraints and maximizes task priority benefits can be selected first; when system resources are tight, a compromise scheme that minimizes resource usage costs and satisfies basic constraints can be selected to achieve efficient resource utilization.

[0117] II. Application Scenarios of Dynamic Tasks

[0118] Taking the intelligent switching of payload observation modes as an example, this paper introduces the intelligent control of dynamic tasks of a geostationary meteorological surface system.

[0119] To address emergency response needs, special support requirements, and user observation requirements for meteorological services, the geostationary meteorological surface system dynamically adjusts the payload observation mode through an intelligent control platform. When an operational observation mode needs to be changed, the system generates the optimal control scheme based on the control type and switching timeliness requirements through an intelligent control matching process, driving multiple systems to collaboratively complete the mode switch. This process includes the following key steps:

[0120] 1) Intelligent control process matching

[0121] The intelligent control management platform generates the optimal execution plan by calling the intelligent control process matching algorithm based on the control type and switching time.

[0122] 2) Task Management and Dispatch Order Issuance

[0123] After receiving the control command, the task management subsystem first generates a new task schedule for the business observation mode and distributes this schedule to each system. It then sends a first-level scheduling order to "update task schedule" and notifies each system to switch to the new business observation mode. Simultaneously, it sends the generated new observation mode schedule to the data archiving and service systems, and broadcasts the schedule to the data acquisition and measurement and control systems.

[0124] 3) Scheduling and Control

[0125] Upon receiving the "Update Task Schedule" scheduling order, each system in the hierarchical scheduling system parses the updated load task schedule, updates the system's job execution plan for that load, and implements data reception, processing, and distribution for the new business observation mode based on the updated job plan.

[0126] The integrated scheduling algorithm adopts the NSGA-III multi-objective evolutionary algorithm. In the triple objective space of time window constraint, maximizing task priority benefits and minimizing execution deviation, the algorithm selects the optimal scheduling scheme in the Pareto front through iterative processes such as population initialization, non-dominated sorting, reference point association and adaptive mutation.

[0127] Among them, the intelligent control process matching algorithm includes three types: intelligent control process A, B, and C (e.g., Figure 2 (As shown). For control flows A and B, the load observation mode switching occurs the following day, and their implementation is relatively simple: mode switching is completed by updating the task template, and system stability can be confirmed through a routine verification process. Although control flow C also adopts a following-day switching strategy, its application window is located after the following-day instruction, increasing its complexity. Its process is shown below. Figure 6 A special case of process C is that the business observation mode needs to be switched over urgently within the current two hours. Its control is the most complex, and its process is as follows: Figure 8 As shown, the following operations need to be completed:

[0128] 1) Send the telemetry and control management command "Pause the telemetry and control of a certain load";

[0129] 2) Send the telemetry and control management command "Clear on-board instructions for a certain payload";

[0130] 3) Switch the task template for the payload's service observation mode;

[0131] 4) Send the task management command "Update and distribute a certain payload task schedule";

[0132] 5) Send the telemetry and control management command "Reset Non-Send Status"

[0133] 6) Send the measurement and control management command "Update a certain load measurement and control plan";

[0134] 7) Check if the command parameter file is updated;

[0135] 8) Send the telemetry and control management command "Resume telemetry and control of a certain load";

[0136] 9) Send the task management command "Issue a certain payload task schedule to all ground systems";

[0137] 10) Determine if the switchover time is before the next day's schedule update is issued. If yes, send an SMS notification that the switchover is complete; otherwise, proceed to step 11.

[0138] 11) Send the task management command "Update and send down the schedule of a certain load task for the next day".

[0139] The above are merely preferred embodiments of the present invention, intended only to aid in understanding the method and core ideas of this application. The scope of protection of the present invention is not limited to the above embodiments; all technical solutions falling within the scope of the present invention's concept are within its protection. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

[0140] This invention comprehensively addresses the shortcomings of existing technologies in intelligent control research and methods for satellite ground control systems facing dynamic tasks. By designing intelligent control algorithms and processes for dynamic tasks, it achieves automatic execution and rapid response of dynamic tasks, effectively meeting the requirements of emergency response, meteorological service support, and satellite safety management. The intelligent control strategy not only improves the system's intelligence level and operational efficiency, providing more precise and efficient services for meteorological observation and satellite operation, but also further reduces the risk of human error.

Claims

1. A static meteorological satellite dynamic task intelligent control method, characterized in that, The method comprises the following steps: S1, generating a dynamic task schedule by modifying a task schedule template, triggering a ground system and its scheduling control by an intelligent control platform driven by the dynamic task schedule; S2, adopting a differentiated scheduling strategy according to characteristics of the dynamic service, adopting integrated scheduling for dynamic service with high timeliness requirement, and adopting hierarchical scheduling for dynamic service with low real-time requirement; S3, designing a scheduling mode and algorithm by deeply analyzing ground task management and control and instruction uploading execution time logic, so as to achieve optimal system performance; S4, according to the dynamic task application time and execution time requirement, the intelligent control platform selects different control processes, specifically: Control flow A: t q ∈ [T0+24h, T0+48h] and t s <T0; Control flow B: t q ∈ [T0+24h, T0+48h] and T0 < t s <T1-Δt_ctrl; Control flow C: t q ∈ [T0+24h, T0+48h] and T1-Δt_ctrl < t s , or t s < T0+24h q On the same day; Wherein, t s is the observation task application time, t q is the task execution time, the system is switched to the dynamic task working mode, T0 is the next day task schedule generation and delivery time; T1 is the next day instruction start loading time, and Δt_ctrl is the minimum control period of the dynamic task. S5, based on the intelligent control process of S4, according to characteristics of the dynamic service, controlling the ground system: for hierarchical scheduling: the ground system distributes the task schedule to each technical system through a message queue, and adopts a standardized scheduling order to issue commands to other systems, the commands include two types of scheduling orders of task schedule arrival and task schedule update, after receiving the task schedule, the technical system needs to feed back a confirmation receipt to the task management and control system, and return communication confirmation information after receiving the scheduling order; for integrated scheduling: the ground system maximizes the task priority benefit f1, minimizes the task execution time deviation f2 and the calculation resource use cost f3, adopts a multi-objective evolutionary algorithm combined with constraint conditions to perform multi-objective optimization, so as to realize global scheduling optimization of the system, and introduces a penalty function to punish individuals that violate the constraint conditions, for guiding the multi-objective evolutionary algorithm to search in the feasible solution direction; S6, calculation of the objective function: Let X = [x1, x2... x N n], where x i ∈{0,1} represents whether task i is scheduled, 1 for scheduled, 0 for not scheduled; Let S = [S1, S2... S N ], denote the start time of each task; Let R = [r1, r2... r N ], represent the resource combination assigned to each task; then the objective function is: Equation One: s.t 0<w i <10; Equation Two: Equation Three: where w i is the weight of task i, t qi is the expected execution time of task i; S7, the multi-objective evolutionary algorithm specifically comprises: S7-1, initializing a population based on a multi-objective optimization model and calculating the objective function value; S7-2, performing hierarchical sorting by non-dominated sorting, generating uniformly distributed reference points and establishing an association relationship between individuals and reference points; S7-3, selecting the next generation population by an elite strategy, and performing crossover and mutation operations on the population; S7-4, repeating S7-1 to S7-3 until the termination condition is met.

2. The intelligent control method for dynamic tasking of geostationary meteorological satellites according to claim 1, characterized in that, The task schedule template comprises an imager task schedule template, a detector task schedule template, a fast imager task schedule template and a satellite orbit determination management task schedule template.

3. The intelligent control method for dynamic tasking of geostationary meteorological satellites according to claim 2, characterized in that, The task schedule template is defined in a task unit and adopts an extensible markup language file format specification, and the format specification comprises a task serial number, a task English name, a task number, a task start time, a task end time and a task parameter.

4. The intelligent control method for dynamic tasking of geostationary meteorological satellites according to claim 1, characterized in that, In S4, T0 is UTC 04:00 every day, T1 is UTC 22:00 every day, and Δt_ctrl is 10 minutes.

5. The intelligent control method for dynamic tasking of geostationary meteorological satellites according to claim 1, characterized in that, The penalty function in the integrated scheduling in S5 specifically comprises: S5-1, constraint on the time window: Start time S of task i i Should be within its allowed time window, i.e.: Equation four: t qi ≤ S i ≤ t qi + Δt window ; where Δt window is the length of the time window for task i, and if task i is scheduled but exceeds the time window, a penalty term is added: Equation Five: t qi Expected execution time for task i S5-2, constraint on the resource: The sum of the resource combinations assigned to the tasks cannot exceed the total resource limit R of the system max That is: Equation Six: If the sum of resources exceeds the limit, add a penalty term: Equation Seven: r i resources allocated to the i-th task; S5-3, constraint on the task dependency relationship: If task j depends on task k, then the end time of task k must be earlier than the start time of task j: Equation Eight: S i ≥ S k + Δt k ; where Δt k is the execution time of task k, and if the dependency is not satisfied, a penalty term is added: Equation Nine: P dependency =∑ 所有依赖关系 max(0,S j -(S k +Δt k ));S k is the execution time of the kth task; S5-4, definition of the comprehensive penalty function: The penalty terms of S5-1-S5-3 are weighted and added to the objective function to obtain a comprehensive penalty function: Formula ten: P = aP + bP + gP time + bP resource + gP dependency ; wherein a, b, g are weight coefficients, which are dynamically adjusted according to the load of the ground system and the dynamic task type, and the number of iterations.

Images