Satellite resource dynamic allocation method and system based on task network business volume prediction

Abstract

The invention discloses a satellite resource dynamic allocation method and system based on task network service volume prediction, and the method specifically comprises the steps: extracting call-related events and resource state information in the operation of a task network, carrying out the preprocessing of the call-related events, and predicting the service volume of the task network according to a time period; when it is detected that the task network carries out service communication resource allocation or a resource allocation period is reached, triggering a resource allocation operation, firstly carrying out information extraction on resource state information, and calculating a difference value between a resource allocation margin of the task network and an idle resource; then calculating a difference value between the predicted resource demand quantity and the current resource quantity in the next resource adjustment period; performing weighted summation on a difference value between the resource configuration allowance and the idle resource and a difference value between the resource prediction demand and the current resource quantity to obtain a resource adjustment parameter; and finally, task network resource adjustment is implemented according to the resource adjustment parameters. According to the invention, the allocation efficiency and utilization efficiency of satellite resources are improved, and the service quality of communication services is improved.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and is a method and system for dynamic allocation of satellite resources based on mission network traffic prediction. Background Technology

[0002] With the continuous breakthroughs in various new communication technologies and the booming development of the satellite internet industry, the technical and application forms of satellite communication systems will undergo revolutionary changes. The increasing abundance of satellite resources, the continuous growth in the number of users, the increasing complexity of constellation networks, and the rapid development of fully flexible high-throughput technologies make resource management increasingly complex. How to efficiently, reliably, and flexibly manage and allocate satellite communication resources to meet the diverse needs of a massive number of users is an urgent task that needs to be carried out and is also the key to maximizing the effectiveness of satellite communication systems.

[0003] Limited by the total amount of communication satellite resources, centralized management and control of satellite resources has become the mainstream application model in the industry. However, centralized management and control restricts the flexible utilization of resources. Many allocation methods and strategies have been researched to solve the problem of dynamic resource adjustment among multi-task networks: some methods focus on resource quality, striving to allocate higher-quality and more usable satellite resources to the task network; some methods focus on network demand, calculating the amount of resources needed to meet the initial service requirements during network construction, and allocating sufficient usable satellite resources to the task network; some methods focus on resource utilization, configuring the limited resources of the task network with more reasonable parameters to achieve higher network service quality. Invention patent CN 115441935 A discloses a situational awareness-based dynamic allocation device for satellite resources in a task network. By monitoring and evaluating the real-time status of network operation and resource availability, it dynamically adjusts the resources used by the satellite communication network, improving resource utilization and the quality of service assurance, enabling satellite resources to achieve greater utilization efficiency. However, these methods are all based on the actual business operations to allocate satellite resources, without considering whether the total amount of satellite resources allocated to the mission network is reasonable and whether it can meet the current and future needs. Furthermore, they do not consider when to implement resource allocation for the mission network to achieve the highest efficiency when the mission network can accept dynamic resource allocation. This is also a key issue that the centralized resource management system needs to address. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for dynamic allocation of satellite resources based on mission network traffic prediction, which features high satellite resource utilization efficiency, high resource allocation efficiency, and a good user experience.

[0005] The technical solution to achieve the purpose of this invention is: a method for dynamic allocation of satellite resources based on mission network traffic prediction, comprising the following steps:

[0006] Step 1: Monitor the status of the normally operating task network in real time and extract call-related events and resource status information during operation;

[0007] Step 2: Format and preprocess call-related events, merge and record them with historical data, and perform data analysis and processing by building a predictive model to predict the traffic volume of the task network according to the time period.

[0008] Step 3: After the task network detects that it has allocated business communication resources, or after the resource allocation cycle has been reached, trigger the resource allocation operation.

[0009] Step 4: Extract information from the resource status information to obtain the idle resource data of the task network, and calculate the difference between the resource configuration margin and the idle resources of the task network.

[0010] Step 5: Calculate the difference between the projected resource demand and the current resource quantity in the next resource adjustment cycle;

[0011] Step 6: Weight the difference between the resource allocation margin and the idle resources of the task network and the difference between the predicted resource demand and the current resource quantity to obtain the resource adjustment parameters;

[0012] Step 7: Implement task network resource adjustments based on resource adjustment parameters.

[0013] Furthermore, the real-time status monitoring of the normally operating task network described in step 1, and the extraction of call-related events and resource status information during operation, are as follows:

[0014] Step 1.1: Monitor the real-time status of the task network, obtain call-related events, and extract call request events, call success events, call rejection events, and call completion events.

[0015] Step 1.2: Obtain the resource status information of the task network and record the resource allocation and resource usage.

[0016] Furthermore, step 2 involves formatting and preprocessing call-related events, merging them with historical data, and then using a predictive model for data analysis to predict the traffic volume of the task network over a time period, as detailed below:

[0017] Step 2.1: Format and preprocess the call-related events extracted in Step 1 to adjust the cycle. To achieve statistical time granularity, the operation of the task network is discretized into each adjustment period. The runtime is calculated and integrated according to the call event time, event type, service demand unit, service type, service rate, and reason code to calculate each adjustment cycle. Internal business resources Specifically:

[0018]

[0019] in, for The amount of service resources successfully called within a given time period. for The amount of service resources that caused call failures within a given time period. These are weighting coefficients, and their values ​​range from [value range missing]. ;

[0020] The calculation formula is:

[0021]

[0022] The calculation formula is:

[0023]

[0024] in, The number of business types. For the first Number of times of similar business, For the first Resource quantity for similar business types;

[0025] Step 2.2: Build a prediction model using machine learning algorithms to predict the task network under different conditions. Business resource volume within the time period , Indicating the future One cycle ;

[0026] Step 2.3: Validate the prediction model using the latest task network monitoring data and fine-tune the model parameters.

[0027] Furthermore, the adjustment cycle described in step 2.1 Setting the value to a smaller value results in higher overall resource utilization efficiency, but also higher requirements for system connectivity and computing power. Therefore, [the value is set to a smaller value]. Set within half an hour.

[0028] Furthermore, after the task network detects the allocation of service communication resources in step 3, or after the resource allocation cycle is reached, a resource allocation operation is triggered, as follows:

[0029] Resource allocation is triggered in two ways: first, by business communication, which occurs after the task network is detected to be allocating business communication resources; and second, by the allocation cycle, which occurs when the resource allocation cycle arrives. The business communication trigger is used to calculate the allocation decision for the task network, while the allocation cycle trigger is used to calculate the allocation decision for all currently managed task networks.

[0030] Further, step 4 involves extracting resource status information to obtain idle resource data for the task network, and calculating the difference between the resource allocation margin and idle resources of the task network, as detailed below:

[0031] Step 4.1: Extract the resource status information from Step 1 to obtain the resource usage of the task network. When allocation is triggered, read the total allocated resources of the task network from the resource management center. Obtain the amount of resources used from the resource usage data of the task network. then the amount of idle resources ;

[0032] Step 4.2: Calculate the resource requirements of the task network, i.e., the difference between the resource allocation margin and the actual idle resources. The resource allocation margin setting will be determined based on the priority of the task network configuration. The resource allocation margin of the task network with higher priority will be set to the amount of resources for the most frequently used service in the task network.

[0033] Furthermore, step 5, which involves calculating the difference between the projected resource demand and the current resource quantity for the next resource adjustment cycle, is detailed as follows:

[0034] Adjust the current resource cycle of the task network. Predicted business resource volume That is, predicting resource demand and comparing it with the current resource allocation of the task network. Subtracting the two yields the difference between the predicted resource demand and the current resource quantity. .

[0035] Furthermore, in step 6, the resource adjustment parameters are obtained by weighted summing the difference between the resource allocation margin and idle resources of the task network and the difference between the predicted resource demand and the current resource quantity, as follows:

[0036] The difference between the task network resource allocation margin obtained in step 3 and the actual idle resources. The difference between the predicted resource demand and the current resource quantity obtained in step 4. Perform a weighted summation to obtain the resource adjustment parameters. The formula is:

[0037]

[0038] in, These are weighting coefficients, and their values ​​range from [value range missing]. .

[0039] Furthermore, step 7, which involves adjusting the task network resources based on the resource adjustment parameters, is detailed as follows:

[0040] Step 7.1, according to Numerical implementation task network resource adjustment, if If the number is negative, it is determined that the task network needs to be reduced. Resources; if If it is zero, no resource adjustment is needed; if If the number must be positive, proceed to step 7.2;

[0041] Step 7.2: Determine if there are any free resources in the resource center. If there are free resources, proceed to step 7.3; otherwise, do not adjust the resources.

[0042] Step 7.3: Determine if there are other high-priority networks that need to be allocated. If so, allocate the high-priority networks first, and return to Step 7.2 to continue determining if there are any idle resources in the resource center. If there are no other high-priority networks that need to be allocated, execute the allocation to increase the task network. Resources.

[0043] A dynamic satellite resource allocation system based on mission network traffic prediction is disclosed. This system implements the aforementioned dynamic satellite resource allocation method based on mission network traffic prediction. The system includes an information extraction module, a traffic prediction module, a resource allocation triggering module, a first difference calculation module, a second difference calculation module, a resource adjustment parameter calculation module, and a resource adjustment module, wherein:

[0044] The information extraction module performs real-time status monitoring of the normally operating task network and extracts call-related events and resource status information during operation.

[0045] The traffic volume prediction module formats and preprocesses call-related events, merges them with historical data, and performs data analysis and processing by building a prediction model to predict the traffic volume of the task network according to the time period.

[0046] The resource allocation trigger module detects that the task network has allocated business communication resources or that the resource allocation cycle has been reached, and then triggers the resource allocation operation.

[0047] The first difference calculation module extracts resource status information to obtain the idle resource data of the task network, and calculates the difference between the resource configuration margin and the idle resources of the task network.

[0048] The second difference calculation module calculates the difference between the predicted resource demand in the next resource adjustment cycle and the current resource quantity.

[0049] The resource adjustment parameter calculation module calculates the resource adjustment parameters by weighting and summing the difference between the resource allocation margin and the idle resources in the task network and the difference between the predicted resource demand and the current resource quantity.

[0050] The resource adjustment module adjusts the task network resources according to the resource adjustment parameters.

[0051] Compared with the prior art, the present invention has the following significant advantages: (1) Based on the actual operation of the network, combined with the prediction of the task network traffic, the service carrying requirements and service guarantee efficiency of the task network are comprehensively considered, thus ensuring the service quality of communication services and improving the utilization efficiency of satellite resources; (2) According to the importance of the network in actual use, a priority scheduling mechanism is set up, thereby ensuring the use of high priority network resources and improving user experience; (3) Satellite resources are allocated according to the service resource allocation and resource allocation cycle, thus improving the efficiency of resource allocation. Attached Figure Description

[0052] Figure 1 This is a schematic diagram illustrating the principle of satellite resource allocation.

[0053] Figure 2 This is a flowchart illustrating a method for dynamic allocation of satellite resources based on mission network traffic prediction, as proposed in this invention. Detailed Implementation

[0054] This invention proposes a dynamic allocation method for satellite resources in a mission network based on traffic volume prediction. This method mainly designs resource allocation strategies through two elements: mission network traffic volume prediction and mission network resource surplus allocation. It completes dynamic resource allocation before the actual occurrence of traffic. On the one hand, it does not mechanically schedule resources based solely on resource utilization thresholds. On the other hand, it ensures that there are available resources when traffic is triggered, thereby guaranteeing both utilization and call connection rates.

[0055] Based on actual application needs, each business communication requires a certain amount of satellite resources. When the task network Triggering business needs, and If there are insufficient available idle resources, the communication request will be rejected.

[0056] The method for dynamic allocation of satellite resources in the mission network based on traffic volume prediction proposed in this invention has the following specific steps:

[0057] Step 1: Perform real-time status monitoring on the normally operating task network, and extract call-related events and resource status information during operation, as follows:

[0058] Step 1.1: Monitor the real-time status of the task network, obtain call-related events, and extract call request events, call success events, call rejection events, and call completion events.

[0059] Step 1.2: Obtain the resource status information of the task network and record the resource allocation and resource usage.

[0060] Step 2: Format and preprocess call-related events, merge and record them with historical data, and perform data analysis by building a predictive model to predict the traffic volume of the task network according to the time period, as follows:

[0061] Step 2.1: Format and preprocess the call-related events extracted in Step 1 to adjust the cycle. To achieve statistical time granularity, the operation of the task network is discretized into each adjustment period. The runtime is calculated and integrated according to the call event time, event type, service demand unit, service type, service rate, and reason code to calculate each adjustment cycle. Internal business resources Specifically:

[0062]

[0063] in, for The amount of service resources successfully called within a given time period. for The amount of service resources that caused call failures within a given time period. These are weighting coefficients, and their values ​​range from [value range missing]. ;

[0064] The calculation formula is:

[0065]

[0066] The calculation formula is:

[0067]

[0068] in, The number of business types. For the first Number of times of similar business, For the first Resource quantity for similar business types;

[0069] Adjustment cycle Setting the value to a smaller value results in higher overall resource utilization efficiency, but also higher requirements for system connectivity and computing power. Therefore, [the value is set to a smaller value]. Set within half an hour.

[0070] Step 2.2: Build a prediction model using machine learning algorithms to predict the task network under different conditions. Business resource volume within the time period , Indicating the future One cycle ;

[0071] Step 2.3: Continuously verify the prediction model with the latest task network monitoring data and fine-tune the model parameters to make the prediction results more accurate.

[0072] Step 3: After the task network detects the allocation of service communication resources, or after the resource allocation cycle is reached, a resource allocation operation is triggered, as follows:

[0073] Resource allocation is triggered in two ways: first, by business communication, which occurs after the task network is detected to be allocating business communication resources; and second, by allocation cycle, which occurs when the resource allocation cycle arrives. The business communication trigger involves calculating the allocation decision for the task network in question, while the allocation cycle trigger involves calculating the allocation decision for all currently managed task networks.

[0074] Step 4: Extract resource status information to obtain the idle resource data of the task network, and calculate the difference between the resource allocation margin and the idle resources of the task network, as detailed below:

[0075] Step 4.1: Extract the resource status information from Step 1 to obtain the resource usage of the task network. When allocation is triggered, read the total allocated resources of the task network from the resource management center. Obtain the amount of resources used from the resource usage data of the task network. then the amount of idle resources ;

[0076] Step 4.2: Calculate the resource requirements of the task network, i.e., the difference between the resource allocation margin and the actual idle resources. The resource allocation margin setting will be determined based on the priority of the task network configuration. The resource allocation margin of a high-priority task network will be set to the amount of resources for the most frequently used service in the task network.

[0077] Step 5: Calculate the difference between the projected resource demand and the current resource quantity in the next resource adjustment cycle, as follows:

[0078] Adjust the current resource cycle of the task network. Predicted business resource volume That is, predicting resource demand and comparing it with the current resource allocation of the task network. Subtracting the two yields the difference between the predicted resource demand and the current resource quantity. .

[0079] Step 6: The resource adjustment parameters are obtained by weighted summing the difference between the resource allocation margin and idle resources in the task network and the difference between the predicted resource demand and the current resource quantity, as follows:

[0080] The difference between the task network resource allocation margin obtained in step 3 and the actual idle resources. The difference between the predicted resource demand and the current resource quantity obtained in step 4. Perform a weighted summation to obtain the resource adjustment parameters. The formula is:

[0081]

[0082] in, These are weighting coefficients, and their values ​​range from [value range missing]. .

[0083] Step 7: Implement task network resource adjustments based on resource adjustment parameters, as detailed below:

[0084] Step 7.1, according to Numerical implementation task network resource adjustment, if If the number is negative, it is determined that the task network needs to be reduced. Resources; if If it is zero, no resource adjustment is needed; if If the number must be positive, proceed to step 7.2;

[0085] Step 7.2: Determine if there are any free resources in the resource center. If there are free resources, proceed to step 7.3; otherwise, do not adjust the resources.

[0086] Step 7.3: Determine if there are other high-priority networks that need to be allocated. If so, allocate the high-priority networks first, and return to Step 7.2 to continue determining if there are any idle resources in the resource center. If there are no other high-priority networks that need to be allocated, execute the allocation to increase the task network. Resources.

[0087] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0088] Example 1

[0089] This invention discloses a method for dynamic allocation of satellite resources based on mission network traffic prediction, comprising the following steps:

[0090] S1. Real-time status monitoring of the task network.

[0091] Monitor the real-time status of the task network, acquire network operation events, and extract call request events, call success events, call rejection events, and call completion events; acquire network resource status and record resource allocation and resource usage.

[0092] S2. Prediction of task network service resource volume.

[0093] The call events in S1 are formatted and preprocessed to adjust the cycle. To achieve statistical time granularity, the operation of the task network is discretized into each... The runtime is calculated and integrated according to the call event time, event type, service demand unit, service type, service rate, reason code, etc., to calculate each Internal business resources .

[0094] The adjustment cycle A smaller setting results in higher overall resource utilization efficiency, but it also places higher demands on system connectivity and computing power. Generally, a smaller setting can be used to optimize resource utilization. Set within half an hour.

[0095] S3. Calculate the resource requirements for the task network configuration.

[0096] Extract the resource status information from S1 to obtain the resource usage of the task network, calculate the amount of idle resources, and calculate the resource configuration requirements of the task network, i.e., the difference between the configured surplus resources and the actual idle resources.

[0097] The resource reserve setting will be determined based on the priority of the task network configuration. The resource reserve of a high-priority task network is usually set to the amount of resources required for the most frequently used service in the task network.

[0098] S4. Calculate resource requirements for the task network prediction.

[0099] The predicted business resource quantity, i.e. the resource demand quantity, of the task network in the current resource adjustment period is obtained from S2. This quantity is then subtracted from the current resource allocation quantity of the task network to calculate the predicted resource demand of the task network.

[0100] S5. Calculation of resource allocation.

[0101] The resource adjustment parameters are obtained by weighted summing of the differences in S3 and S4. When calculating the allocation amount, partial data from both types of resources are used, i.e., a weighted sum is performed. The weighting value can be adjusted to cope with different emergency scenarios.

[0102] S6. Real-time task allocation based on resource allocation volume.

[0103] Resource allocation is performed based on the calculation results of S5, including three cases: positive, negative, and zero, which correspond to resource increase, decrease, and no change, respectively.

[0104] When making adjustments, the priority attributes of the task network and the remaining resources in resource management should be taken into account.

[0105] Step S2 further includes: This prediction mode mainly predicts the future through historical data, that is, it constructs a time-series-based business volume distribution relationship through historical business volume data of the task network, obtains business volume prediction data within a scheduling cycle, and continuously verifies the prediction model and optimizes the model parameters with the latest task network monitoring data to make the prediction results more accurate.

[0106] Step S6 further includes: setting priorities for task networks, and when resource increases are triggered, sorting the task networks that need to be increased from high to low priority, and allocating the available resources of the resource center to higher-level tasks in a priority manner.

[0107] The resource allocation method has two triggering times: one is after the task network is detected to be allocating business communication resources, and the other is when the resource allocation cycle arrives.

[0108] Example 2

[0109] In this embodiment, the satellite resource allocation is as follows: Figure 1 As shown, the resource management center centrally manages satellite resources and allocates resources to different networks based on the mission network situation.

[0110] This invention proposes a method for dynamic allocation of task network resources based on traffic volume forecasting, namely, enabling the resource management center to dynamically allocate task network resources based on traffic volume forecasting. The process is as follows: Figure 2 As shown, the steps are as follows:

[0111] S1. Real-time status monitoring of the task network.

[0112] Monitor the real-time status of the task network, acquire network operation events, and extract call request events, call success events, call rejection events, and call completion events; acquire network resource status and record resource allocation and resource usage.

[0113] S2. Prediction of task network service resource volume.

[0114] The call events in S1 are formatted and preprocessed to adjust the cycle. To achieve statistical time granularity, the operation of the task network is discretized into each... The runtime is calculated and integrated according to the call event time, event type, service demand unit, service type, service rate, reason code, etc., to calculate each Internal business resources Specifically

[0115]

[0116] in, for The amount of service resources successfully called within a given time period. for The amount of service resources for call failures within a given time period is a weighting factor, with a value range of [value missing]. In practical applications, the equation

[0117]

[0118] This should be true, but considering that call request events and call success / rejection events may be separated into two adjacent statistical periods, the number of call requests is not directly used as the business demand. In addition, the reasons for call rejection generally include: incorrect calling party information, the called party not being on the network, insufficient resources, etc. For call rejections caused by incorrect calling party information, there are often subsequent calls after the relevant information is corrected. That is, the number of call rejections can reflect a certain degree of real business demand. Therefore, the number of call rejections is weighted in the calculation.

[0119] The calculation formula is:

[0120]

[0121] in, The number of business types. For the first Number of times of similar business, For the first Resource quantity for similar business types.

[0122] The calculation formula is:

[0123]

[0124] The meaning is the same as above.

[0125] The predictive model is built using machine learning algorithms to predict the task network under different conditions. Business resource volume within the time period , Indicating the future One cycle Meanwhile, the latest task network monitoring data is continuously used to verify the prediction model and optimize the model parameters to make the prediction results more accurate.

[0126] The adjustment cycle A smaller setting results in higher overall resource utilization efficiency, but it also places higher demands on system connectivity and computing power. Generally, a smaller setting can be used to optimize resource utilization. Set within half an hour.

[0127] S3. Calculate the resource requirements for the task network configuration.

[0128] Data is extracted from the resource status information in S1 to obtain the resource usage of the task network. When allocation is triggered, the total allocated resources of the task network are read from the resource management center. Obtain the amount of resources used from the resource usage data of the task network. then the amount of idle resources .

[0129] Calculate the resource requirements for the task network configuration, i.e., the difference between the configured surplus resources and the actual idle resources. The resource reserve setting will be determined based on the priority of the task network configuration. The resource reserve of a high-priority task network is usually set to the amount of resources required for a single operation of the most frequently used service in the task network.

[0130] S4. Calculate resource requirements for the task network prediction.

[0131] Obtain the current resource adjustment cycle of the task network from S2. Predicted business resource volume That is, the resource demand and the current resource allocation of the task network. Subtracting these values, we can calculate the predicted resource requirements of the task network, i.e., the difference between the predicted resources and the allocated resources. .

[0132] S5. Calculation of resource allocation.

[0133] The difference in S3 The difference between S4 and S4 Perform a weighted summation to obtain the resource adjustment parameters. Specifically

[0134]

[0135] in, These are weighting coefficients, and their values ​​range from [value range missing]. As historical data accumulates, prediction models become increasingly accurate. However, the timing of sudden business events lacks regularity. Therefore, this invention addresses this issue by setting resource reserves. When calculating allocation quantities, two types of resource demands are weighted and summed, meaning partial data from both types of resources are considered. The weighting values ​​are adjustable; different weights can address different sudden scenarios. For example, if the regularity of business triggering is low, the accuracy of business prediction is considered low, and the data from resource reserves is given more weight during calculation. The smaller the value, the better, and vice versa.

[0136] S6. Real-time task allocation based on resource allocation volume.

[0137] according to Numerical implementation task network resource adjustment, when If the number is negative, it is determined that the task network needs to be reduced. Resources; when If the value is zero, no resource adjustment is needed; when If the number is positive, it is considered an increase in the task network. Resources. In the case of resource reduction, allocation can be performed immediately, while the allocation decision for resource increase must be stored in the allocation execution buffer.

[0138] Resource allocation is triggered in two ways: first, by business communication, which occurs after the task network is detected allocating business communication resources; and second, by the allocation cycle, which occurs when the resource allocation cycle arrives. The business communication trigger primarily involves calculating the allocation decision for that specific task network, while the allocation cycle trigger involves calculating the allocation decision for all currently managed task networks.

[0139] In summary, S1 and S2 of the method are in normal operation after the system starts. When resource allocation is triggered, S3 to S6 are executed sequentially on the relevant task networks. After the allocation decision results of the task networks are entered into the execution buffer, they are sorted and processed according to the priority of the task networks. That is, under the premise that there are enough spare resources in the resource management center, the allocation of high-priority task networks is executed first.

[0140] Example 3

[0141] In this embodiment, a satellite resource dynamic allocation system based on mission network traffic prediction is also provided. This system is used to implement the aforementioned satellite resource dynamic allocation method based on mission network traffic prediction. The system includes an information extraction module, a traffic prediction module, a resource allocation triggering module, a first difference calculation module, a second difference calculation module, a resource adjustment parameter calculation module, and a resource adjustment module, wherein:

[0142] The information extraction module performs real-time status monitoring of the normally operating task network and extracts call-related events and resource status information during operation.

[0143] The traffic volume prediction module formats and preprocesses call-related events, merges them with historical data, and performs data analysis and processing by building a prediction model to predict the traffic volume of the task network according to the time period.

[0144] The resource allocation trigger module detects that the task network has allocated business communication resources or that the resource allocation cycle has been reached, and then triggers the resource allocation operation.

[0145] The first difference calculation module extracts resource status information to obtain the idle resource data of the task network, and calculates the difference between the resource configuration margin and the idle resources of the task network.

[0146] The second difference calculation module calculates the difference between the predicted resource demand in the next resource adjustment cycle and the current resource quantity.

[0147] The resource adjustment parameter calculation module calculates the resource adjustment parameters by weighting and summing the difference between the resource allocation margin and the idle resources in the task network and the difference between the predicted resource demand and the current resource quantity.

[0148] The resource adjustment module adjusts the task network resources according to the resource adjustment parameters.

[0149] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for dynamic allocation of satellite resources based on mission network traffic prediction, characterized in that, Includes the following steps: Step 1: Monitor the status of the normally operating task network in real time and extract call-related events and resource status information during operation; Step 2: Format and preprocess call-related events, merge and record them with historical data, and perform data analysis and processing by building a predictive model to predict the traffic volume of the task network according to the time period. Step 3: After the task network detects that it has allocated business communication resources, or after the resource allocation cycle has been reached, trigger the resource allocation operation. Step 4: Extract information from the resource status information to obtain the idle resource data of the task network, and calculate the difference between the resource configuration margin and the idle resources of the task network. Step 5: Calculate the difference between the projected resource demand and the current resource quantity in the next resource adjustment cycle; Step 6: Weight the difference between the resource allocation margin and the idle resources of the task network and the difference between the predicted resource demand and the current resource quantity to obtain the resource adjustment parameters; Step 7: Implement task network resource adjustments based on resource adjustment parameters.

2. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 1, characterized in that, Step 1 involves real-time status monitoring of the normally operating task network, extracting call-related events and resource status information during operation, as detailed below: Step 1.1: Monitor the real-time status of the task network, obtain call-related events, and extract call request events, call success events, call rejection events, and call completion events. Step 1.2: Obtain the resource status information of the task network and record the resource allocation and resource usage.

3. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 2, characterized in that, Step 2 involves formatting and preprocessing call-related events, merging them with historical data, and then using a predictive model for data analysis to predict the traffic volume of the task network over a time period. The details are as follows: Step 2.1: Format and preprocess the call-related events extracted in Step 1 to adjust the cycle. To achieve statistical time granularity, the operation of the task network is discretized into each adjustment period. The runtime is calculated and integrated according to the call event time, event type, service demand unit, service type, service rate, and reason code to calculate each adjustment cycle. Internal business resources Specifically: ； in, for The amount of service resources successfully called within a given time period. for The amount of service resources that caused call failures within a given time period. These are weighting coefficients, and their values ​​range from [value range missing]. ; The calculation formula is: ； The calculation formula is: ； in, The number of business types. For the first Number of times of similar business, For the first Resource quantity for similar business types; Step 2.2: Build a prediction model using machine learning algorithms to predict the task network under different conditions. Business resource volume within the time period , Indicating the future One cycle ; Step 2.3: Validate the prediction model using the latest task network monitoring data and fine-tune the model parameters.

4. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 3, characterized in that, The adjustment cycle described in step 2.1 Setting the value to a smaller value results in higher overall resource utilization efficiency, but also higher requirements for system connectivity and computing power. Therefore, [the value is set to a smaller value]. Set within half an hour.

5. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 4, characterized in that, Step 3, which involves detecting that the task network is allocating service communication resources or reaching the resource allocation cycle, triggers a resource allocation operation, as detailed below: Resource allocation is triggered in two ways: first, by business communication, which occurs after the task network is detected to be allocating business communication resources; and second, by the allocation cycle, which occurs when the resource allocation cycle arrives. The business communication trigger is used to calculate the allocation decision for the task network, while the allocation cycle trigger is used to calculate the allocation decision for all currently managed task networks.

6. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 5, characterized in that, Step 4 involves extracting resource status information to obtain idle resource data for the task network, and calculating the difference between the task network's resource allocation margin and idle resources, as detailed below: Step 4.1: Extract the resource status information from Step 1 to obtain the resource usage of the task network. When allocation is triggered, read the total allocated resources of the task network from the resource management center. Obtain the amount of resources used from the resource usage data of the task network. then the amount of idle resources ; Step 4.2: Calculate the resource requirements of the task network, i.e., the difference between the resource allocation margin and the actual idle resources. The resource allocation margin setting will be determined based on the priority of the task network configuration. The resource allocation margin of the task network with higher priority will be set to the amount of resources for the most frequently used service in the task network.

7. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 6, characterized in that, Step 5, which involves calculating the difference between the projected resource demand and the current resource quantity for the next resource adjustment cycle, is as follows: Adjust the current resource cycle of the task network. Predicted business resource volume That is, predicting resource demand and comparing it with the current resource allocation of the task network. Subtracting the two yields the difference between the predicted resource demand and the current resource quantity. .

8. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 7, characterized in that, Step 6 involves weighted summation of the difference between the resource allocation margin and idle resources in the task network and the difference between the predicted resource demand and the current resource quantity to obtain the resource adjustment parameters, as detailed below: The difference between the task network resource allocation margin obtained in step 3 and the actual idle resources. The difference between the predicted resource demand and the current resource quantity obtained in step 4. Perform a weighted summation to obtain the resource adjustment parameters. The formula is: ； in, These are weighting coefficients, and their values ​​range from [value range missing]. .

9. The method for dynamic allocation of satellite resources based on mission network traffic prediction according to claim 8, characterized in that, Step 7, which describes adjusting the task network resources based on the resource adjustment parameters, is as follows: Step 7.1, according to Numerical implementation task network resource adjustment, if If the number is negative, it is determined that the task network needs to be reduced. Resources; if If it is zero, no resource adjustment is needed; if If the number must be positive, proceed to step 7.2; Step 7.2: Determine if there are any free resources in the resource center. If there are free resources, proceed to step 7.3; otherwise, do not adjust the resources. Step 7.3: Determine if there are other high-priority networks that need to be allocated. If so, allocate the high-priority networks first, and return to Step 7.2 to continue determining if there are any idle resources in the resource center. If there are no other high-priority networks that need to be allocated, execute the allocation to increase the task network. Resources.

10. A dynamic satellite resource allocation system based on mission network traffic prediction, characterized in that, This system is used to implement the dynamic satellite resource allocation method based on mission network traffic prediction as described in any one of claims 1 to 9. The system includes an information extraction module, a traffic prediction module, a resource allocation triggering module, a first difference calculation module, a second difference calculation module, a resource adjustment parameter calculation module, and a resource adjustment module, wherein: The information extraction module performs real-time status monitoring of the normally operating task network and extracts call-related events and resource status information during operation. The traffic volume prediction module formats and preprocesses call-related events, merges them with historical data, and performs data analysis and processing by building a prediction model to predict the traffic volume of the task network according to the time period. The resource allocation trigger module detects that the task network has allocated business communication resources or that the resource allocation cycle has been reached, and then triggers the resource allocation operation. The first difference calculation module extracts resource status information to obtain the idle resource data of the task network, and calculates the difference between the resource configuration margin and the idle resources of the task network. The second difference calculation module calculates the difference between the predicted resource demand in the next resource adjustment cycle and the current resource quantity. The resource adjustment parameter calculation module calculates the resource adjustment parameters by weighting and summing the difference between the resource allocation margin and the idle resources in the task network and the difference between the predicted resource demand and the current resource quantity. The resource adjustment module adjusts the task network resources according to the resource adjustment parameters.

Images