A 5g network capacity and energy saving collaborative control method and system

Abstract

The application discloses a 5G network capacity and energy saving collaborative control method and system, and relates to the technical field of communication networks.The method comprises the following steps: collecting multi-source heterogeneous data of a target area, including historical performance data in the network and external event data; inputting the multi-source heterogeneous data into a preset prediction model to output a service load prediction result of each cell in each time period of the target area in a future time period; determining an expansion time period and an energy saving time period according to the prediction result, generating an expansion strategy for the expansion time period, and generating an energy saving strategy for the energy saving time period; and issuing the expansion strategy and the energy saving strategy to corresponding network devices for execution.The application realizes accurate allocation of network resources and energy efficiency optimization by fusing tidal prediction of multi-source heterogeneous data and bidirectional collaborative planning of capacity and energy saving, and solves the problem that it is difficult to balance experience guarantee and energy consumption reduction in a tidal scene caused by the fragmentation of expansion and energy saving.

Description

Technical Field

[0001] This invention relates to the field of communication network technology, and in particular to a method and system for coordinated control of 5G network capacity and energy saving. Background Technology

[0002] With the large-scale deployment of 5G networks, various cultural and tourism scenic spots, transportation hubs, commercial centers and other areas are showing obvious business tidal characteristics. Network traffic surges sharply during peak hours and drops significantly during off-peak hours. This poses dual requirements for network capacity scheduling and base station energy consumption management. How to flexibly allocate network resources according to changes in business load and balance service capacity and operating power consumption has become a key issue that needs to be addressed in the field of 5G intelligent operation and maintenance.

[0003] Currently, the industry generally adopts base station hibernation technology based on load thresholds and static time rules to cope with network tidal changes. By monitoring core indicators such as cell PRB utilization, energy-saving operations such as carrier frequency and radio frequency channel shutdown are triggered when the load is lower than a preset threshold. When the load increases, the base station resumes normal operation. This method achieves preliminary control of base station energy consumption based on simple threshold judgment, and reduces power consumption during network idle periods to a certain extent.

[0004] However, this type of passive response management method has obvious shortcomings. It not only relies on historical data within the network to judge changes in services, but also has difficulty predicting sudden traffic fluctuations caused by external factors such as holidays and large-scale events. The prediction accuracy is insufficient. More importantly, it separates capacity expansion and base station energy saving into two independent management processes. The two cannot work together, which ultimately leads to the network either sacrificing energy saving to avoid congestion or excessive energy saving affecting the user's communication experience during peak hours. It is difficult to achieve the optimal synergy between capacity guarantee and energy saving. Summary of the Invention

[0005] This invention provides a method and system for coordinated control of 5G network capacity and energy saving, which can solve the problems of fragmented capacity and energy saving strategies, single prediction dimension, and lack of closed-loop feedback in the prior art. It realizes bidirectional coordination of capacity expansion and hierarchical energy saving based on multi-source prediction, as well as quantitative evaluation of strategy execution and model self-evolution.

[0006] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions: Firstly, a method for coordinated control of 5G network capacity and energy saving is provided. The method includes: collecting multi-source heterogeneous data of a target area, the multi-source heterogeneous data including historical performance data within the network and external event data; inputting the multi-source heterogeneous data into a preset prediction model, and outputting a service load prediction result for the target area in a future time period, the service load prediction result including the predicted service load index of each cell in each time period; determining a first time period and a second time period in the future time period based on the service load prediction result, wherein the predicted service load of the first time period meets the capacity expansion trigger condition and the predicted service load of the second time period meets the energy saving trigger condition; generating a capacity expansion strategy for the first time period and an energy saving strategy for the second time period; and distributing the capacity expansion strategy and the energy saving strategy to the corresponding network devices for execution.

[0007] The method provided by this invention overcomes the limitation of existing technologies that rely solely on intranet KPIs for prediction by collecting and integrating multi-source heterogeneous data, including historical performance data within the network and external event data. Based on the same service load prediction results, it simultaneously determines the first time period that meets the expansion conditions and the second time period that meets the energy-saving conditions, and generates expansion strategies and energy-saving strategies in parallel and in a targeted manner. This solves the contradiction in existing technologies where expansion and energy saving are handled separately, making it difficult to balance "ensuring experience and reducing energy consumption" in tidal scenarios.

[0008] In one possible implementation of the first aspect, the intranet historical performance data includes at least one of PRB utilization, RRC connection count, and user plane throughput; the external event data includes at least one of holiday calendar data, large-scale event announcement data, and social media popularity data.

[0009] In one possible implementation of the first aspect, the preset prediction model is a hybrid neural network model of LSTM and attention mechanism.

[0010] The method provided by this invention uses a hybrid architecture of LSTM and attention mechanism to adaptively extract time-series features and external event features that are strongly correlated with business load prediction from multi-source time-series data, thereby improving the prediction accuracy of sudden tidal traffic caused by holidays, large-scale events and other events.

[0011] In one possible implementation of the first aspect, the capacity expansion triggering condition is: the predicted PRB utilization rate of the target cell in the target time period is greater than a first utilization rate threshold and the predicted number of users is greater than a first number of users threshold; the energy saving triggering condition is: the predicted PRB utilization rate of the target cell in the target time period is less than a second utilization rate threshold.

[0012] The method provided by this invention ensures that capacity expansion is triggered only when user experience is truly needed, and energy saving is enabled only when the load is truly reduced to a safe level, by setting dual expansion trigger conditions for PRB utilization and number of users, as well as energy-saving trigger conditions for PRB utilization.

[0013] In one possible implementation of the first aspect, the capacity expansion strategy includes: applying for a carrier aggregation license from the license pool and activating the secondary carrier during the first time period; the energy saving strategy includes: performing at least one of carrier frequency shutdown, channel shutdown, or deep sleep during the second time period, with the energy saving intensity dynamically adjusted according to the predicted service load depth of the second time period.

[0014] The method provided by this invention achieves flexible expansion of software-defined capacity through dynamic allocation of license pools and activation of carrier aggregation. It achieves refined energy consumption management through graded energy-saving means such as carrier frequency shutdown, channel shutdown, or deep sleep. The energy-saving intensity is dynamically adjusted according to the predicted load depth, avoiding the drawbacks of fixed threshold energy saving.

[0015] In one possible implementation of the first aspect, the validity period of the expansion strategy is aligned with the start and end times of the first time period; the execution time window of the energy-saving strategy is aligned with the start and end times of the second time period.

[0016] The method provided by this invention strictly aligns the capacity expansion validity period and energy-saving time window with the corresponding prediction period, ensuring that capacity expansion resources are only occupied during peak business periods and energy-saving strategies only take effect during off-peak business periods, thereby achieving precise allocation of network resources and optimal management of energy efficiency.

[0017] In one possible implementation of the first aspect, the method further includes: pre-setting multiple scenario strategy templates to a strategy rule library, each scenario strategy template containing expansion rules and energy-saving rules for a corresponding scenario; when the business load prediction result matches the target scenario strategy template in the strategy rule library, calling the target scenario strategy template to generate the expansion strategy and the energy-saving strategy.

[0018] The method provided by this invention pre-configures a scenario strategy template library and encapsulates expert experience into reusable rule templates. When a corresponding scenario is matched, the generated strategy is directly invoked, which reduces the deployment threshold in complex scenarios and improves the generalization ability and automation of the solution.

[0019] In one possible implementation of the first aspect, the scenario strategy template includes a scenic area daytime peak template, a venue instantaneous peak template, or a transportation hub morning peak template.

[0020] In one possible implementation of the first aspect, the step of sending the capacity expansion strategy and the energy-saving strategy to the corresponding network device for execution includes: sending the capacity expansion strategy to the gNodeB through the northbound interface to trigger license loading and carrier activation; writing the energy-saving strategy into the base station energy-saving task scheduler and automatically starting and stopping it within a preset time window; performing legality verification on the sent strategy instructions, triggering an automatic alarm and retaining the original configuration when execution fails.

[0021] The method provided by this invention realizes the automated distribution of policies through the northbound interface. The expansion and energy-saving instructions are executed through different channels. While ensuring that the instructions take effect accurately, it also has the ability to verify the validity and roll back in case of failure, thus ensuring the reliability and security of the automated execution process.

[0022] In one possible implementation of the first aspect, the method further includes: after the strategy execution is completed, collecting actual performance data and actual energy consumption data, comparing the actual performance data with the business load prediction result, comparing the actual energy consumption data with the baseline energy consumption, and obtaining a quantitative evaluation result of the strategy execution effect; feeding the quantitative evaluation result back to the preset prediction model and strategy rule base for model iteration and strategy parameter optimization.

[0023] The method provided by this invention compares and evaluates the actual KPIs after strategy execution with the predicted values ​​and the actual energy consumption with the baseline energy consumption. The evaluation results are fed back to the prediction model and the strategy rule base to drive continuous iterative optimization, which solves the problem that existing open-loop control schemes lack self-evolution capabilities.

[0024] In one possible implementation of the first aspect, the quantitative evaluation results include prediction accuracy, energy saving rate, and user experience assurance indicators; the prediction accuracy is calculated based on the deviation between the actual performance data and the business load prediction results, and the energy saving rate is calculated based on the difference between the baseline energy consumption and the actual energy consumption.

[0025] In one possible implementation of the first aspect, when the prediction accuracy is less than a preset accuracy threshold, the preset prediction model retraining process is triggered; when the energy-saving strategy causes performance degradation after execution, the corresponding safety margin parameter in the strategy rule base is adjusted.

[0026] The method provided by this invention achieves automated closed-loop tuning of system parameters by setting a prediction accuracy threshold to trigger model retraining and automatically adjusting the safety margin by monitoring performance degradation caused by energy saving.

[0027] Secondly, the present invention provides a 5G network capacity and energy-saving collaborative control system, the system comprising: a multi-source data fusion prediction module, used to collect multi-source heterogeneous data of a target area, input the multi-source heterogeneous data into a preset prediction model, and output the service load prediction result of the target area in a future time period; a joint planning module, used to determine a first time period and a second time period in the future time period based on the service load prediction result, generate a capacity expansion strategy in parallel for the first time period, and generate an energy-saving strategy in parallel for the second time period; a strategy execution engine, used to distribute the capacity expansion strategy and the energy-saving strategy to the corresponding network devices for execution; and a closed-loop evaluation module, used to collect actual performance data and actual energy consumption data after the strategy execution is completed, compare the actual performance data with the service load prediction result, compare the actual energy consumption data with the baseline energy consumption, obtain a quantitative evaluation result of the strategy execution effect, and feed the quantitative evaluation result back to the preset prediction model and strategy rule base.

[0028] Thirdly, an electronic device is provided, the electronic device including a memory and one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, which, when executed by the processor, cause the electronic device to perform the method as described in any implementation of the first aspect.

[0029] Fourthly, a computer-readable storage medium is provided, including computer instructions that, when executed on an electronic device, cause the electronic device to perform a method as described in any implementation of the first aspect.

[0030] Understandably, the beneficial effects of the system of the second aspect, the electronic device of the third aspect, and the computer-readable storage medium of the fourth aspect provided above can be referenced to the beneficial effects of the first aspect and any of its possible design embodiments, which will not be repeated here. Attached Figure Description

[0031] Figure 1 An overall architecture diagram of a 5G network capacity and energy-saving collaborative control system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention; Figure 3 A flowchart illustrating a 5G network capacity and energy-saving coordinated control method provided in an embodiment of the present invention; Figure 4 This is a diagram illustrating a closed-loop evaluation feedback mechanism provided in an embodiment of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be described below with reference to the accompanying drawings. In the description of the present invention, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. The "or" in the present invention is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A or B can represent: A alone, A and B simultaneously, and B alone. A and B can be singular or plural. Furthermore, in the description of the present invention, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items.

[0033] Furthermore, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" are not necessarily different.

[0034] In this embodiment of the invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in this embodiment of the invention should not be construed as superior or more advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0035] With the large-scale deployment of 5G networks, various cultural and tourism scenic spots, transportation hubs, commercial centers and other areas are showing obvious business tidal characteristics. Network traffic surges sharply during peak hours and drops significantly during off-peak hours. This poses dual requirements for network capacity scheduling and base station energy consumption management. How to flexibly allocate network resources according to changes in business load and balance service capacity and operating power consumption has become a key issue that needs to be addressed in the field of 5G intelligent operation and maintenance.

[0036] Currently, the industry generally adopts base station hibernation technology based on load thresholds and static time rules to cope with network tidal changes. By monitoring core indicators such as cell PRB utilization, energy-saving operations such as carrier frequency and radio frequency channel shutdown are triggered when the load is lower than a preset threshold. When the load increases, the base station resumes normal operation. This method achieves preliminary control of base station energy consumption based on simple threshold judgment, and reduces power consumption during network idle periods to a certain extent.

[0037] However, this type of passive response management has obvious shortcomings. It not only relies on historical data within the network to judge changes in services, but also has difficulty predicting sudden traffic fluctuations caused by external factors such as holidays and large-scale events. The prediction accuracy is insufficient. More importantly, it separates capacity expansion and base station energy saving into two independent management processes. The two cannot work together. At the same time, there is a lack of quantitative evaluation and a closed-loop mechanism for model self-evolution after the strategy is implemented. Ultimately, the network either sacrifices energy saving to avoid congestion or excessive energy saving affects the user communication experience during peak hours. It is difficult to achieve the optimal synergy between capacity guarantee and energy saving.

[0038] In view of this, embodiments of the present invention provide a 5G network capacity and energy-saving coordinated control method, comprising: collecting multi-source heterogeneous data of a target area, the multi-source heterogeneous data including intra-network historical performance data and external event data; inputting the multi-source heterogeneous data into a preset prediction model, and outputting a service load prediction result of the target area in a future time period, the service load prediction result including the predicted service load index of each cell in each time period; determining a first time period and a second time period in the future time period based on the service load prediction result, wherein the predicted service load of the first time period meets the capacity expansion trigger condition and the predicted service load of the second time period meets the energy-saving trigger condition; generating a capacity expansion strategy for the first time period and an energy-saving strategy for the second time period; and distributing the capacity expansion strategy and the energy-saving strategy to the corresponding network devices for execution.

[0039] The method provided by this invention integrates multi-source heterogeneous data input to the prediction model by fusing historical performance data within the network and external event data. Based on the same prediction result, it synchronously generates two types of strategies: capacity expansion and energy saving, thus achieving bidirectional coordination between capacity guarantee and energy efficiency optimization in tidal scenarios.

[0040] In some embodiments, a 5G network capacity and energy-saving coordinated control method provided in this invention can be executed by a 5G network capacity and energy-saving coordinated control system (hereinafter referred to as coordinated control system 100).

[0041] See Figure 1 , Figure 1This invention provides a system architecture diagram. The collaborative control system 100 includes: a multi-source data fusion prediction module 110, a joint planning module 120, a strategy execution engine 130, and a closed-loop evaluation module 140. Data flow is divided into two categories: the prediction process involves multi-source data input to the fusion prediction module 110, which outputs service load prediction results; the joint planning module 120 receives the prediction results and simultaneously outputs expansion and energy-saving strategies; the strategy execution engine 130 interfaces with both the wireless network element (gNodeB) and the energy-saving scheduling platform, executing precisely within a specified time window; the closed-loop evaluation module 140 automatically collects actual KPIs and energy consumption data after an event, calculates evaluation indicators by comparing predicted values ​​with actual performance, and sends feedback information back to the prediction model and strategy rule base. The entire process is fully automatic and uninterrupted, forming a complete closed-loop system.

[0042] As an example, the collaborative control system 100 can be any electronic device 200 with data processing capabilities, such as a server, a general-purpose computer, or a cloud platform. The specific implementation method of the collaborative control system is not limited here.

[0043] Figure 2 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention is shown. The electronic device 200 includes a processor 210, a memory 220, and a communication interface 230.

[0044] Processor 210 may include one or more processing cores. Processor 210 connects to various parts within electronic device 200 using various interfaces and lines, and performs various functions and processes data of electronic device 200 by running or executing instructions, programs, code sets, or instruction sets stored in memory 220, and by calling data stored in memory 220. Optionally, processor 210 may be implemented using at least one of the following hardware forms: Central Processing Unit (CPU), Graphics Processing Unit (GPU), Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA).

[0045] The memory 220 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 220 may include a non-transitory computer-readable storage medium. The memory 220 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 220 may include a program storage area. This program storage area may store instructions for implementing an operating system, instructions for implementing at least one function, instructions for implementing the various method embodiments described above, etc.

[0046] Communication interface 230 is used to communicate with other devices, equipment or communication networks, such as data storage devices, image processing devices or Ethernet, wireless access network (RAN), wireless local area network (WLAN), etc.

[0047] In terms of physical implementation, the aforementioned devices (such as processor 210, memory 220, and communication interface 230) can each be devices within the same device (such as a laptop computer). Alternatively, at least two of these devices can be located within the same device, i.e., as different devices within the same device, similar to the deployment of devices or components in a distributed system.

[0048] It is understood that the structure illustrated in this embodiment does not constitute a specific limitation on the electronic device 200. In other embodiments of the present invention, the electronic device 200 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0049] The following description, in conjunction with the accompanying drawings, illustrates a method for coordinated control of 5G network capacity and energy saving provided by an embodiment of the present invention.

[0050] Figure 3 This is a flowchart illustrating a 5G network capacity and energy-saving coordinated control method provided in an embodiment of the present invention. Optionally, this method can be... Figure 1 The illustrated electronic device 200 performs this method, which includes the following steps: S1. Collect multi-source heterogeneous data of the target area, including historical performance data within the network and external event data.

[0051] Specifically, the target areas are 5G network coverage areas with significant business fluctuations, such as cultural and tourism scenic spots, large commercial districts, sports venues, and transportation hubs—places with high population mobility and drastic, periodic fluctuations in business volume. These areas share the common characteristic of distinct busy and off-peak periods: during peak hours, users are highly concentrated, generating extremely high concurrent business demands, while during off-peak hours, few users lead to significant idle network resources.

[0052] The system synchronously collects multi-source heterogeneous data from multiple data sources, breaking through the limitations of traditional methods that rely solely on a single data source of network KPIs. The multi-source heterogeneous data is divided into two main categories based on its source: historical performance data from within the network and external event data. These two categories differ fundamentally in their collection methods and data structures, requiring separate processing and fusion.

[0053] Historical performance data within the network is obtained from the operator's OSS / BSS system. OSS (Operation Support System) provides network-side performance statistics, while BSS (Business Support System) provides user-side service statistics. In one possible implementation, historical performance data within the network includes at least one of PRB utilization, RRC connections, and user plane throughput. PRB utilization reflects the occupancy of radio resource blocks in the time-frequency domain and is a core indicator for measuring the busy / idle status of cell radio resources; RRC connections reflect the number of user devices in the connected state within the cell and is a basic indicator for measuring the scale of access users; user plane throughput reflects the total uplink and downlink data traffic actually generated by all users within the cell and is a direct indicator for measuring service load.

[0054] The system can collect historical data from the same period over the past few months to years as training samples for the model. For example, for the prediction task of the National Day Golden Week, the system collects hourly PRB utilization, RRC connection number and user plane throughput data of each cell in the target area during the National Day period (October 1 to October 7) over the past three years as the basic features for time series modeling.

[0055] External event data is acquired from external information sources and is used to capture macro-environmental factors related to business load but not observable from internal network data. In one possible implementation, external event data includes at least one of the following: holiday calendar data, large-scale event announcement data, and social media trend data. Holiday calendar data includes national statutory holiday arrangements, work schedule adjustments, and traditional festival dates, used to label the time attributes of different dates; large-scale event announcement data includes information such as daily performances, expected visitor flow, and opening hours released by scenic spots, venues, and other public channels, used to quantify the expected scale of crowd gathering in a specific area; social media trend data includes quantitative indicators reflecting the current social attention to a specific area or event, such as Weibo topic index, Douyin check-in volume, and short video posting frequency, used to capture sudden changes in trending topics that traditional statistical data cannot reflect.

[0056] It should be noted that the above multi-source data were cleaned and formatted after collection to eliminate format differences, unit differences, and time alignment deviations between different data sources, and were unified into a structured data format indexed by cell identifier and time period, providing a standardized input basis for subsequent model prediction.

[0057] S2. Input the multi-source heterogeneous data into a preset prediction model and output the service load prediction result of the target area in the future time period. The service load prediction result includes the predicted service load index of each cell in each time period.

[0058] In one possible implementation, the pre-defined prediction model is a hybrid neural network model combining LSTM and an attention mechanism. LSTM (Long Short-Term Memory) is a recurrent neural network structure designed for time-series data. Through gating mechanisms such as forget gates, input gates, and output gates, it can effectively capture long-term dependencies and periodic patterns in time series, avoiding the gradient vanishing problem that traditional recurrent neural networks encounter when processing long sequences. The attention mechanism allows the model to dynamically adjust the degree of attention given to different dimensions of features when processing input features, assigning higher attention weights to features more strongly correlated with the current prediction target.

[0059] In practical applications, this hybrid model aligns the feature dimensions of multi-source heterogeneous data temporally as the input multidimensional feature vector. The LSTM network layer is responsible for mining the temporal variation patterns of business load, including daytime cycles (daytime peaks, nighttime troughs), weekday cycles (differences between weekdays and weekends), and annual cycles (differences between peak and off-peak seasons). The attention mechanism layer is responsible for adaptively adjusting the weights of different features: during normal periods, the model mainly focuses on the temporal variation patterns of historical performance data within the network; during holidays and large-scale events, the model automatically increases the attention weight of external event features (such as social media popularity, event announcements, and visitor flow) to capture the additional impact of external events on business load.

[0060] The model outputs predicted service load metrics for each cell in the target area over a future period, including predicted PRB utilization, predicted RRC connections, and predicted user plane throughput. In one possible implementation, the prediction period is the next 7 days, output in hourly granularity, resulting in predicted values ​​for 168 consecutive time points.

[0061] For example, the LSTM+Attention hybrid neural network model has been validated in a live network pilot test, with a prediction error of less than 10% during the National Day Golden Week, which is significantly better than the traditional CNN-LSTM model and time-series smoothing prediction algorithm.

[0062] S3. Based on the business load prediction results, determine the first time period and the second time period in the future time period. The predicted business load of the first time period meets the capacity expansion trigger condition, and the predicted business load of the second time period meets the energy saving trigger condition.

[0063] The system reads the hourly service load prediction results output by the prediction model, and determines the expansion trigger conditions and energy-saving trigger conditions for each cell in each hour of the future time period. The time periods that meet the conditions are marked as expansion periods or energy-saving periods, respectively.

[0064] In one possible implementation, the expansion trigger condition is: the predicted PRB utilization rate of the target cell during the target time period is greater than a first utilization rate threshold and the predicted number of users is greater than a first user number threshold. For example, the first utilization rate threshold can be set to 85%, and the first user number threshold can be set separately according to the capacity specifications of each cell. This dual judgment mechanism of PRB utilization rate and user number ensures that expansion is only triggered when the traffic volume truly approaches the network limit, avoiding misjudgment based on a single indicator. For example, if the predicted PRB utilization rate of a cell near a scenic area reaches 88% and the predicted number of RRC connections exceeds 120% of the cell's rated capacity during the time period of 14:00 on October 2nd, then this time period meets the expansion trigger condition and is marked as the first time period.

[0065] The energy-saving trigger condition is: the predicted PRB utilization rate of the target cell during the target time period is less than a second utilization rate threshold. For example, the second utilization rate threshold can be set to 30%. The reason for using only PRB utilization rate as the energy-saving criterion is that a low PRB utilization rate itself indicates a large amount of idle wireless resources, eliminating the need for additional reference to user count indicators. For example, if the predicted PRB utilization rate of a cell near the same scenic area is only 8% from 02:00 to 06:00 on October 3rd, this continuous period meets the energy-saving trigger condition and is marked as the second time period.

[0066] It should be noted that the first and second time periods originate from different time segments of the same forecast result and may be located at different times in the future. A community may simultaneously have expansion periods and energy-saving periods, which do not overlap and correspond to different time windows and different strategic actions.

[0067] S4. Generate a capacity expansion strategy for the first time period and a power saving strategy for the second time period.

[0068] In one possible implementation, the capacity expansion strategy includes: applying for a carrier aggregation license from the license pool and activating the secondary carrier during the first time period. The license pool is a software license resource pool centrally managed by the operator, containing various software licenses such as carrier aggregation licenses and MIMO function licenses, supporting dynamic allocation on demand and timely recovery. Based on the predicted capacity expansion demand, the system automatically applies for a carrier aggregation license valid for a specified time period from the license pool, automatically loads the license and activates the secondary carrier at the start of the first time period, increasing the cell peak rate and capacity to the rated level under carrier aggregation mode. The validity period of the capacity expansion strategy is strictly aligned with the start and end times of the first time period. After the capacity expansion period ends, the license resources are automatically released and returned to the license pool, achieving flexible resource management of "precise deployment and immediate return after use".

[0069] The energy-saving strategy includes at least one of carrier frequency shutdown, channel shutdown, or deep sleep during the second time period, with the energy-saving intensity dynamically adjusted according to the predicted service load depth during the second time period. Carrier frequency shutdown involves turning off some carrier frequency units, reducing the power consumption of the radio frequency module while preserving basic coverage capabilities; channel shutdown involves further shutting down some radio frequency channels, suitable for scenarios with extremely low predicted load; deep sleep involves putting most functional modules other than the base station core control module into a sleep state to achieve maximum energy-saving effect.

[0070] In one possible implementation, the energy-saving intensity tiered matching rule is as follows: when the predicted PRB utilization is between 20% and 30%, carrier frequency shutdown is executed; when the predicted PRB utilization is between 10% and 20%, a combined strategy of carrier frequency shutdown and channel shutdown is executed; when the predicted PRB utilization is below 10%, deep sleep is enabled. The execution time window of the energy-saving strategy is strictly aligned with the start and end times of the second time period.

[0071] Meanwhile, the collaborative control system 100 supports rapid strategy generation based on scenario templates. In one possible implementation, the system pre-defines multiple scenario strategy templates in the strategy rule library. Each scenario strategy template encapsulates expansion rules and energy-saving rules for a corresponding typical scenario. Scenario strategy templates include a daytime peak template for scenic spots (adapted to the characteristics of peak daytime visitor flow and rapid decline at night in cultural and tourism scenic spots), a venue instantaneous surge template (adapted to the characteristics of large-scale gatherings of people in short periods such as sports venues and concerts), or a morning peak template for transportation hubs (adapted to the characteristics of peak commuting periods such as train stations and airports). When the temporal characteristics and peak load patterns of the business load prediction results match the target scenario strategy template in the strategy rule library, the system directly calls the expansion rules and energy-saving rules in that template to generate the corresponding expansion strategy and energy-saving strategy, without needing to calculate trigger conditions and strategy parameters item by item, significantly improving the efficiency of strategy generation.

[0072] S5. The expansion strategy and the energy-saving strategy are sent to the corresponding network devices for execution.

[0073] In one possible implementation, the policy distribution is divided into two independent execution channels: capacity expansion and energy saving.

[0074] For capacity expansion, the expansion strategy is distributed to the gNodeB via the northbound interface. The northbound interface is a standardized management interface between the OSS / network management system and the wireless network element, supporting protocols such as RESTful API and CORBA. The expansion command includes a license loading request and a carrier activation command, triggering the gNodeB to automatically activate the secondary carrier at the start of the first time period, increasing the peak rate and capacity of the cell to the target level. Before distribution, the system verifies the legality of the strategy command, including whether the current remaining amount in the license pool meets the requested number of licenses, whether the target base station's software version supports carrier aggregation, and whether the target cell's hardware configuration has secondary carrier resources. If any verification item fails, an automatic alarm is triggered and the process is redirected to manual review, while the original cell configuration is preserved to ensure that the network operation is not interrupted due to erroneous commands.

[0075] In terms of energy conservation, energy-saving strategies are incorporated into the base station energy-saving task scheduler. The base station energy-saving task scheduler is a task scheduling module built into the network management system or the base station side, supporting automatic execution of energy-saving actions according to preset time windows. The energy-saving strategy automatically triggers the corresponding shutdown action at the beginning of the second time period and automatically restores the base station to normal operating status at the end of the second time period. Similarly, before issuing energy-saving commands, a validity check is performed, including whether the shutdown action conflicts with other base station maintenance tasks and whether the coverage after shutdown meets the minimum continuity requirements. If execution fails, an automatic alarm is triggered, and the original configuration is retained.

[0076] The above-mentioned capacity expansion and energy-saving strategies can be implemented without human intervention, and the entire process is automated.

[0077] In some embodiments, the method provided by the present invention further includes: After the strategy execution is completed, actual performance data and actual energy consumption data are collected. The actual performance data is compared with the business load prediction results, and the actual energy consumption data is compared with the baseline energy consumption to obtain a quantitative evaluation result of the strategy execution effect. The quantitative evaluation result is fed back to the preset prediction model and strategy rule base for model iteration and strategy parameter optimization.

[0078] Specifically, after the policy execution concludes (e.g., the day after the National Day Golden Week ends), the closed-loop evaluation module collects the following data from the OSS and energy management platform during the policy execution period: First, actual performance data, namely, the actual PRB utilization rate, RRC connection count, and user plane throughput of each cell during the expansion and energy-saving periods, aggregated on an hourly basis; Second, actual energy consumption data, namely, the actual power consumption of each base station, aggregated on an hourly basis. The baseline energy consumption is the historical average energy consumption during the second period when the energy-saving policy is not implemented, used as a reference benchmark for calculating the energy-saving effect.

[0079] In one possible implementation, the quantitative evaluation results include prediction accuracy, energy saving rate, and user experience assurance metrics.

[0080] Prediction accuracy is calculated based on the deviation between actual performance data and the predicted business load, using the following formula: Prediction accuracy = 1 - |actual PRB - predicted PRB| / predicted PRB. When the prediction accuracy is greater than 90%, the prediction quality is considered to be up to standard.

[0081] The energy saving rate is calculated based on the difference between baseline energy consumption and actual energy consumption, using the following formula: Energy saving rate = (baseline energy consumption - actual energy consumption) / baseline energy consumption. When the energy saving rate is greater than 30%, the energy saving effect is considered to meet the standard.

[0082] User experience assurance metrics include the disconnection rate (target less than 0.5%) and average user speed (target greater than 50Mbps) during policy implementation, used to verify that the energy-saving policy does not have a significant impact on user perception.

[0083] In one possible implementation, the closed-loop feedback mechanism works as follows: when the prediction accuracy is less than a preset accuracy threshold (e.g., 90%), a pre-defined retraining process for the prediction model is triggered. The actual data collected in this round is used as new training samples to retrain the model parameters and improve prediction accuracy for similar scenarios. When the energy-saving strategy causes performance degradation (e.g., excessive disconnection rate or substandard average user speed), the corresponding safety margin parameters in the strategy rule base are adjusted. For example, the threshold for the second utilization rate is increased from 30% to 35%, or the threshold for PRB utilization rate, which is judged to be in deep sleep, is decreased from 10% to 5%, sacrificing some energy-saving effects to ensure user experience. All evaluation results are stored in the model training database for the next prediction and planning optimization.

[0084] See Figure 4 , Figure 4 This diagram illustrates a closed-loop evaluation and feedback mechanism provided in an embodiment of the present invention. The diagram shows that after the strategy execution is complete, the system automatically collects key performance indicators and energy consumption data (KPIs) from actual operation. By comparing the predicted values ​​with the actual performance, it calculates evaluation indicators such as prediction accuracy and energy efficiency (energy saving rate), and feeds this information back to the model training database and the strategy rule base for further optimization of prediction model parameters and adjustment of threshold parameters in the strategy rules.

[0085] As can be seen from the above, the method provided by the embodiments of the present invention collects multi-source heterogeneous data of historical performance data and external event data within the fusion network, uses an LSTM+Attention hybrid neural network model to output service load prediction results, generates expansion strategies and hierarchical energy-saving strategies in parallel based on the same prediction results, issues them for execution through the northbound interface and the energy-saving scheduler respectively, and performs closed-loop evaluation and feedback after the strategy ends, transforming expansion and energy saving from two separate independent processes into collaborative decision-making based on the same prediction benchmark.

[0086] This invention can also be understood as follows: It adopts a multi-source heterogeneous data prediction model that integrates in-network performance KPIs with external data such as holidays, event announcements, and social media popularity, breaking through the limitation of existing technologies that rely solely on in-network KPIs. Based on the same prediction result, it constructs a bidirectional collaborative generation mechanism for capacity expansion strategies and tiered energy-saving strategies, with the validity periods of capacity expansion and energy saving strictly aligned with the prediction time periods, thus resolving the contradiction between capacity assurance and energy efficiency optimization in tidal scenarios. It establishes a quantitative evaluation and feedback loop after strategy execution, and achieves continuous self-evolution of the system through prediction accuracy monitoring, performance degradation detection, and automatic adjustment of safety margin.

[0087] The foregoing mainly describes the solutions of the embodiments of the present invention from a methodological perspective. It is understood that, in order to achieve the above-mentioned functions, the cooperative control system 100 includes at least one of the hardware structures and software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, the present invention can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of the present invention.

[0088] In alternative implementations, the LSTM+Attention hybrid neural network model can be replaced with other deep learning models with time series prediction capabilities, such as Transformer, Temporal Fusion Transformer (TFT), Prophet algorithm, or Graph Neural Network (GNN). As long as it can integrate multi-source time series data and external event features to achieve load prediction, it falls within the protection scope of this invention.

[0089] In addition to dynamic license allocation, capacity expansion methods can be extended to the elastic scaling of vCU / vDU virtual resources in cloud-based radio access networks (Cloud RAN) environments, or to achieve capacity elasticity through dynamic adjustment of network slice bandwidth. Energy-saving strategies, in addition to carrier frequency shutdown, channel shutdown, and deep sleep, can further introduce symbol-level shutdown as a more granular energy-saving control action.

[0090] In addition to PRB utilization and absolute energy consumption, evaluation metrics such as user-perceived throughput (50th / 95th percentile throughput), dropout rate, and video stuttering rate can be added as inputs for evaluation feedback.

[0091] This invention also provides a computer-readable storage medium storing at least one computer instruction, which is loaded and executed by a processor to implement the methods of the various embodiments described above. Explanations of the relevant content and descriptions of the beneficial effects of any of the computer-readable storage media provided above can be found in the corresponding embodiments described above, and will not be repeated here.

[0092] This invention also provides a chip. This chip integrates a control circuit for implementing the functions of the aforementioned cooperative control system 100 and one or more ports. Optionally, the functions supported by this chip are as described above and will not be repeated here.

[0093] Those skilled in the art will understand that the program for implementing all or part of the steps of the above embodiments, which can be executed by a program instructing related hardware, can be stored in a computer-readable storage medium. The storage medium mentioned above can be a read-only memory, a random access memory, etc. The processing unit or processor mentioned above can be a central processing unit, a general-purpose processor, an application-specific integrated circuit (ASIC), a microprocessor (DSP), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof.

[0094] This invention also provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform any of the methods described in the above embodiments. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., SSD), etc.

[0095] It should be noted that the devices for storing computer instructions or computer programs provided in the embodiments of the present invention, such as, but not limited to, the aforementioned memory, computer-readable storage medium, and communication chip, are all non-transitory. Those skilled in the art should recognize that the functions described in the embodiments of the present invention in one or more of the above examples can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium accessible to general-purpose or special-purpose computers.

[0096] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for coordinated control of 5G network capacity and energy saving, characterized in that, include: Collect multi-source heterogeneous data of the target area, including historical performance data within the network and external event data; The multi-source heterogeneous data is input into a preset prediction model, and the service load prediction results of the target area in the future time period are output. The service load prediction results include the predicted service load indicators of each cell in each time period. Based on the business load forecast results, a first time period and a second time period in the future time period are determined. The forecast business load of the first time period meets the capacity expansion trigger condition, and the forecast business load of the second time period meets the energy saving trigger condition. A capacity expansion strategy is generated for the first time period, and an energy-saving strategy is generated for the second time period. The expansion strategy and the energy-saving strategy are distributed to the corresponding network devices for execution.

2. The method according to claim 1, characterized in that, The network-internal historical performance data includes at least one of PRB utilization, RRC connection count, and user plane throughput; the external event data includes at least one of holiday calendar data, large-scale event announcement data, and social media popularity data.

3. The method according to claim 1, characterized in that, The preset prediction model is a hybrid neural network model of LSTM and attention mechanism.

4. The method according to claim 1, characterized in that, The capacity expansion trigger condition is: the predicted PRB utilization rate of the target cell in the target time period is greater than the first utilization rate threshold and the predicted number of users is greater than the first number of users threshold; the energy saving trigger condition is: the predicted PRB utilization rate of the target cell in the target time period is less than the second utilization rate threshold.

5. The method according to claim 4, characterized in that, The capacity expansion strategy includes: applying for a carrier aggregation license from the license pool and activating the secondary carrier during the first time period; the energy saving strategy includes: performing at least one of carrier frequency shutdown, channel shutdown, or deep sleep during the second time period, with the energy saving intensity dynamically adjusted according to the predicted service load depth during the second time period.

6. The method according to claim 5, characterized in that, The validity period of the capacity expansion strategy is aligned with the start and end times of the first time period; the execution time window of the energy-saving strategy is aligned with the start and end times of the second time period.

7. The method according to claim 1, characterized in that, The method further includes: Multiple scenario strategy templates are pre-set to the strategy rule library, and each scenario strategy template contains expansion rules and energy-saving rules for the corresponding scenario; When the business load prediction result matches the target scenario policy template in the policy rule base, the target scenario policy template is invoked to generate the expansion policy and the energy saving policy.

8. The method according to claim 7, characterized in that, The scenario strategy templates include daytime peak templates for scenic spots, instantaneous peak templates for venues, or morning peak templates for transportation hubs.

9. The method according to claim 1, characterized in that, The step of distributing the capacity expansion strategy and the energy-saving strategy to the corresponding network devices for execution includes: The expansion strategy is sent to the gNodeB via the northbound interface, triggering license loading and carrier activation; The energy-saving strategy is written into the base station energy-saving task scheduler and automatically started and stopped within a preset time window; The issued policy commands are validated for legality. If execution fails, an automatic alarm is triggered and the original configuration is retained.

10. The method according to claim 7, characterized in that, The method further includes: After the strategy execution is completed, actual performance data and actual energy consumption data are collected. The actual performance data is compared with the business load prediction results, and the actual energy consumption data is compared with the baseline energy consumption to obtain a quantitative evaluation result of the strategy execution effect. The quantitative evaluation results are fed back to the preset prediction model and strategy rule base for model iteration and strategy parameter optimization.

11. The method according to claim 10, characterized in that, The quantitative evaluation results include prediction accuracy, energy saving rate, and user experience assurance indicators; the prediction accuracy is calculated based on the deviation between the actual performance data and the business load prediction results, and the energy saving rate is calculated based on the difference between the baseline energy consumption and the actual energy consumption.

12. The method according to claim 10, characterized in that, When the prediction accuracy is less than a preset accuracy threshold, the preset prediction model retraining process is triggered; when the energy-saving strategy causes performance degradation, the corresponding safety margin parameter in the strategy rule base is adjusted.

13. A 5G network capacity and energy-saving collaborative control system, characterized in that, include: The multi-source data fusion prediction module is used to collect multi-source heterogeneous data of the target area, input the multi-source heterogeneous data into a preset prediction model, and output the business load prediction result of the target area in the future time period. The joint planning module is used to determine the first time period and the second time period in the future based on the business load prediction results, generate a capacity expansion strategy in parallel for the first time period, and generate an energy saving strategy in parallel for the second time period. A policy execution engine is used to distribute the capacity expansion policy and the energy saving policy to the corresponding network devices for execution. The closed-loop evaluation module is used to collect actual performance data and actual energy consumption data after the strategy execution is completed, compare the actual performance data with the business load prediction results, compare the actual energy consumption data with the baseline energy consumption, obtain a quantitative evaluation result of the strategy execution effect, and feed the quantitative evaluation result back to the preset prediction model and strategy rule base.

14. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the 5G network capacity and energy-saving coordinated control method as described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which is loaded and executed by a processor to enable the computer to implement the 5G network capacity and energy-saving coordinated control method as described in any one of claims 1-12.

Images