Cell parameter adaptive adjustment method and device, equipment and storage medium

Abstract

The application provides a cell parameter self-adaptive adjustment method, device, equipment and storage medium, and relates to the technical field of communication. The method comprises the following steps: setting discrete feasible adjustment values based on cell physical parameters in each scheduling period, superimposing the current cell physical parameters and the corresponding feasible adjustment values to obtain the adjustment action of each physical parameter, and combining the adjustment actions of various physical parameters to generate a cell candidate adjustment action set; constructing a cell target utility function based on multiple target parameters; calculating the target utility value after the execution of each candidate adjustment action in the cell candidate adjustment action set; selecting the candidate adjustment action with the maximum target utility value from the cell candidate adjustment action set as the optimal adjustment action under the premise of meeting the preset constraint condition; and executing the optimal adjustment action in the current scheduling period. The application can realize the active perception of network state and the multi-parameter collaborative self-adaptive optimization.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a cell parameter adaptive adjustment method, a cell parameter adaptive adjustment device, a computer device, and a computer-readable storage medium. Background Technology

[0002] In cellular mobile communication systems (such as LTE and NR), base stations (such as eNodeB and gNodeB) are the core nodes responsible for network coverage, capacity, and user experience. The coverage area, signal quality, and inter-cell load balancing of a base station directly depend on its radio parameter settings. In cellular networks, the management of cell boundaries and resource allocation directly impacts the user experience (especially for edge users) and the overall system capacity.

[0003] In existing mobile network optimization efforts, operators primarily adjust wireless parameters by manually analyzing KPI (Key Performance Indicators) and DT (Drive Test) / CQT (Call Quality Test) data, combined with experience.

[0004] However, existing methods for adjusting cell parameters lack proactive response and multi-parameter collaborative optimization capabilities. Summary of the Invention

[0005] This invention was developed to at least partially address the technical problems of existing cell parameter adjustment methods, such as the lack of proactive response and multi-parameter collaborative optimization capabilities.

[0006] According to one aspect of the present invention, a cell parameter adaptive adjustment method is provided, comprising the following steps repeatedly executed in each scheduling cycle: Discrete feasible adjustment values ​​are set based on the cell physical parameters, and the current cell physical parameters are superimposed with the corresponding feasible adjustment values ​​to obtain the adjustment action for each physical parameter. Then, the adjustment actions of various physical parameters are combined to generate a set of candidate cell adjustment actions. The physical parameters include at least one of cell transmit power, antenna downtilt angle and cell individual offset. A target utility function for a cell is constructed based on multiple target parameters; wherein, the target parameters involve at least two of network performance, user experience, handover overhead, and energy consumption. Based on the cell target utility function, calculate the target utility value after the execution of each candidate adjustment action in the cell candidate adjustment action set; Under the premise of satisfying preset constraints, the candidate adjustment action with the largest target utility value is selected from the set of candidate adjustment actions for the cell as the optimal adjustment action; The optimal adjustment action is executed in the current scheduling cycle.

[0007] Optionally, the target utility function of the cell is as follows: ; in, Let c be the target utility function; Let be the set of users connected to community c; Preset throughput weights; The throughput of user u after the candidate adjustment action is executed; To prevent a zero coefficient; Preset user experience weights; The user experience quality metrics after the candidate adjustment action is executed; Preset switching penalty weights; This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Preset energy consumption penalty weights; This represents the change in energy consumption of cell c after the candidate adjustment action is executed.

[0008] Optionally, the formula for calculating the throughput of user u after the candidate adjustment action is executed is as follows: ; in, The throughput of user u after the candidate adjustment action is executed; The bandwidth allocated to user u; The SINR value of the user after the candidate adjustment action is executed; ; in, The SINR value of the user after the candidate adjustment action is executed; The RSRP value of the user after the candidate adjustment action is executed; This represents the interference power received by user u from neighboring cell i at the current moment. This is the sum of the interference power received by user u from all neighboring cells of cell c at the current moment; Preset noise power; ; in, The RSRP value of user u after the candidate adjustment action is executed; The adjusted transmit power of cell c after the candidate adjustment action is executed; for Path loss at that point, and The distance from user u to the base station; This refers to the antenna gain in the user's u direction at the adjusted antenna downtilt angle in cell c after the candidate adjustment action is executed.

[0009] Optionally, the formula for calculating the user u's experience quality index after the candidate adjustment action is executed is as follows: ; in, The user experience quality metrics after the candidate adjustment action is executed; For rate weighting coefficients; The predicted downlink rate for user u after the candidate adjustment action is executed; This is the rate normalization parameter; This is the time delay weighting coefficient; This is the predicted service latency value for user u after the candidate adjustment action is executed; This is the time delay normalization parameter.

[0010] Optionally, the formula for calculating the total number of handovers by users in cell c after the candidate adjustment action is executed is as follows: ; in, This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Let be the set of users connected to cell c; 1{·} is an indicator function that takes the value 1 when the condition is true, and 0 otherwise; The RSRP value received by user u from cell c after the candidate adjustment action is executed; The individual offset of cell c after the candidate adjustment action is executed; The RSRP value received by user u from neighboring cell i at the current moment; This represents the individual offset of neighboring cell i at the current moment.

[0011] Optionally, the formula for calculating the change in energy consumption of cell c after the candidate adjustment action is executed is as follows: ; in, This represents the change in energy consumption of cell c after the candidate adjustment action is executed. The energy consumption value of cell c after the candidate adjustment action is executed; This represents the energy consumption value of cell c at the current moment. Preset RF power efficiency factor; The adjusted transmit power of cell c after the candidate adjustment action is executed; The load rate of cell c after the candidate adjustment action is executed; This represents the transmit power of cell c at the current moment; This represents the load rate of cell c at the current moment.

[0012] Optionally, the step of selecting the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions for the cell, under the premise of satisfying preset constraints, as the optimal adjustment action, includes: From the set of candidate adjustment actions for the cell, candidate adjustment actions that meet the preset constraints are selected as valid candidate adjustment actions, and constitute a subset of valid candidate adjustment actions; Select valid candidate adjustment actions that meet the preset amplitude limit conditions from the subset of valid candidate adjustment actions, and use them as feasible adjustment actions to form a subset of feasible adjustment actions; Select the feasible adjustment action with the largest target utility value from the subset of feasible adjustment actions as the optimal adjustment action.

[0013] Optionally, the preset constraints include: Coverage constraints: ;in, Adjust the action for the i-th candidate The RSRP value of user u after execution; The preset lower limit of signal strength; Switching count constraint: ;in, For the i-th candidate adjustment action The total number of handovers by users within cell c after execution; This is the preset maximum number of switching attempts; Physical parameter limitations: ; ; ; in, This is the preset lower limit of the cell's transmit power; The cell transmit power corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit of cell transmit power; This is the preset lower limit of the antenna downtilt angle; The antenna downtilt angle corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit of the antenna downtilt angle; This is the preset lower limit for individual offsets within the community; The cell offset corresponding to the optimal adjustment action executed in the previous cycle; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset within the community.

[0014] Optionally, the preset amplitude limiting conditions include: ; ; ; in, For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit for power adjustment; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit for antenna downtilt adjustment; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset adjustment within the community.

[0015] According to another aspect of the present invention, a cell parameter adaptive adjustment device is provided, comprising the following modules that work collaboratively in each scheduling cycle: The candidate adjustment action generation module is configured to set discrete feasible adjustment values ​​based on cell physical parameters, and superimpose the current cell physical parameters with the corresponding feasible adjustment values ​​to obtain the adjustment action for each physical parameter. Then, the adjustment actions of various physical parameters are combined to generate a set of candidate cell adjustment actions. The physical parameters include at least one of cell transmit power, antenna downtilt angle, and cell individual offset. The target utility function construction module is configured to construct a cell target utility function based on multiple target parameters; wherein, the target parameters involve at least two of network performance, user experience, handover overhead, and energy consumption; The target utility value calculation module is configured to calculate the target utility value after the execution of each candidate adjustment action in the set of candidate adjustment actions for the cell, based on the target utility function of the cell. The optimal adjustment action determination module is configured to select the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions of the cell, under the premise of satisfying preset constraints, as the optimal adjustment action; The optimal adjustment action execution module is configured to execute the optimal adjustment action in the current scheduling period.

[0016] According to another aspect of the present invention, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the aforementioned cell parameter adaptive adjustment method.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, wherein when the computer program is executed by a processor, the processor performs the aforementioned cell parameter adaptive adjustment method.

[0018] The technical solution provided by this invention may include the following beneficial effects: The cell parameter adaptive adjustment method and apparatus provided by this invention form a closed-loop periodic self-optimizing cell parameter adjustment mechanism by repeatedly executing the generation of a candidate adjustment action set, the calculation of the target utility value, the screening of candidate adjustment actions under constraints, and the execution of the optimal adjustment action in each scheduling cycle. This enables cell parameters to be dynamically adjusted according to network conditions. Moreover, this periodic proactive response method can proactively adjust cell physical parameters according to real-time network conditions and achieve coordinated optimization among multiple parameters such as transmit power, antenna downtilt angle, and individual cell offset. This solves the problem of lack of proactive response and multi-parameter coordinated optimization capabilities in parameter adjustment in the prior art, effectively improving the network's adaptability to sudden traffic and changes in user distribution, while also improving edge user experience and overall network performance.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims, and drawings. Attached Figure Description

[0020] The accompanying drawings are provided to further understand the technical solutions of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the technical solutions of the present invention, and do not constitute a limitation on the technical solutions of the present invention.

[0021] Figure 1 A flowchart illustrating a cell parameter adaptive adjustment method provided in an embodiment of the present invention; Figure 2 A flowchart illustrating another cell parameter adaptive adjustment method provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the cell parameter adaptive adjustment device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the specific implementation methods of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific implementation methods described herein are for illustration and explanation only and are not intended to limit the present invention.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a set order or sequence; furthermore, in the absence of conflict, the embodiments and features in the embodiments of this invention can be arbitrarily combined with each other. In the following description, the use of suffixes such as "module," "component," or "unit" to represent elements is only for the convenience of the description of this invention and has no inherent meaning. Therefore, "module," "component," or "unit" can be used interchangeably.

[0024] In related technologies, the cell parameter adjustment methods have the following drawbacks: The parameter adjustment process is based on offline optimization using historical data, which cannot respond in a timely manner to sudden changes in traffic (such as concerts, holidays, or a surge in pedestrian traffic in shopping districts). The inability to predict future user movement trends and switching behaviors means that parameter adjustments are often reactive rather than proactive. Typically, only a single parameter is adjusted (e.g., only power or downtilt angle is adjusted), ignoring the coupling effect between multiple parameters, resulting in limited optimization effect or even mutual cancellation. The parameters are often adjusted in fixed steps (such as power ±2 dB), failing to dynamically determine the optimal adjustment range according to the scenario, resulting in insufficient refinement.

[0025] To address the aforementioned problems, this invention provides a cell parameter adaptive adjustment scheme that combines user mobility trend prediction, real-time network status assessment, and multi-parameter collaborative optimization to achieve rapid response to changes in traffic and user distribution, thereby improving edge user experience and overall network performance. Specific embodiments are described in detail below.

[0026] Figure 1 This is a flowchart illustrating a cell parameter adaptive adjustment method provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S101 to S105, which are repeatedly executed in each scheduling cycle.

[0027] S101. Based on the physical parameters of the cell, set discrete feasible adjustment values, and superimpose the current physical parameters of the cell with the corresponding feasible adjustment values ​​to obtain the adjustment action for each physical parameter. Then, combine the adjustment actions of various physical parameters to generate a set of candidate adjustment actions for the cell.

[0028] The physical parameters include at least one of cell transmit power, antenna downtilt angle, and cell individual offset.

[0029] S102. Construct a cell target utility function based on multiple target parameters.

[0030] The target parameters involve at least two of network performance, user experience, switching overhead, and energy consumption.

[0031] S103. Based on the cell target utility function, calculate the target utility value after each candidate adjustment action in the cell candidate adjustment action set is executed.

[0032] S104. Under the premise of satisfying the preset constraints, select the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions for the cell as the optimal adjustment action.

[0033] S105. Execute the optimal adjustment action in the current scheduling cycle.

[0034] In this embodiment, a closed-loop periodic self-optimizing cell parameter adjustment mechanism is formed by repeatedly executing the generation of a candidate adjustment action set, the calculation of the target utility value, the screening of candidate adjustment actions under constraints, and the execution of the optimal adjustment action in each scheduling cycle. This enables cell parameters to be dynamically adjusted according to the network status. Moreover, this periodic proactive response method can proactively adjust cell physical parameters according to the real-time network status and achieve coordinated optimization among multiple parameters such as transmit power, antenna downtilt angle, and individual cell offset. This solves the problem of lack of proactive response and multi-parameter coordinated optimization capabilities in parameter adjustment in the prior art, effectively improving the network's adaptability to sudden traffic and changes in user distribution, while also improving edge user experience and overall network performance.

[0035] In one specific implementation, the cell target utility function constructed in step S102 is as follows: .

[0036] in, Let c be the target utility function; Let be the set of users connected to community c; Preset throughput weights; The throughput of user u after the candidate adjustment action is executed; To prevent a zero coefficient; Preset user experience weights; The user experience quality metrics after the candidate adjustment action is executed; Preset switching penalty weights; This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Preset energy consumption penalty weights; This represents the change in energy consumption of cell c after the candidate adjustment action is executed.

[0037] In this embodiment, a multi-objective weighted utility function is adopted, which includes throughput, quality of experience, number of handovers, and energy consumption changes. This comprehensively balances the conflicts between different optimization objectives and avoids side effects such as a surge in handovers or excessive energy consumption caused by focusing on a single indicator. In addition, the weight coefficients can be flexibly configured according to the operator's strategy, so that the optimization results are more in line with actual operation and maintenance needs, thereby achieving a more balanced and stable improvement in network performance.

[0038] In one specific implementation, the formula for calculating the throughput of user u after the candidate adjustment action is executed is as follows: ; in, The throughput of user u after the candidate adjustment action is executed; The bandwidth allocated to user u; The SINR value of the user after the candidate adjustment action is executed.

[0039] In this embodiment, the Shannon formula is used to predict user throughput after the execution of candidate adjustment actions, which can directly map the improvement of physical layer signal-to-interference-plus-noise ratio (SINR) to the improvement of data link layer transmission capacity. This formula reflects the marginal decrease of throughput with increasing SINR in logarithmic form, avoiding the prediction bias caused by linear assumptions, and making the throughput term evaluation in the target utility function closer to the actual capacity characteristics of the wireless channel.

[0040] In one specific implementation, the formula for calculating the user's SINR value after the candidate adjustment action is executed is as follows: .

[0041] in, The SINR value of the user after the candidate adjustment action is executed; The RSRP value of the user after the candidate adjustment action is executed; This represents the interference power received by user u from neighboring cell i at the current moment. This is the sum of the interference power received by user u from all neighboring cells of cell c at the current moment; This is the preset noise power.

[0042] In this embodiment, the serving signal power after the candidate adjustment action is executed is combined with the neighboring cell interference power at the current moment, which enables the estimation of the SINR value after the candidate adjustment action is executed without additional prediction of interference changes. This formula makes reasonable use of the engineering fact that interference power is relatively stable in the short term, avoids the complexity and uncertainty introduced by predicting future interference, and provides sufficient accuracy while ensuring real-time prediction.

[0043] In one specific implementation, the formula for calculating the user's RSRP value after the candidate adjustment action is executed is as follows: .

[0044] in, The RSRP value of user u after the candidate adjustment action is executed; The adjusted transmit power of cell c after the candidate adjustment action is executed; for Path loss at that location, and The distance from user u to the base station; This refers to the antenna gain in the user's u direction at the adjusted antenna downtilt angle in cell c after the candidate adjustment action is executed.

[0045] In this embodiment, the three factors of transmit power, path loss, and antenna gain are integrated into one, enabling the quantification of the combined impact of transmit power adjustment and antenna downtilt adjustment on the user's received signal strength in candidate actions. Specifically, the antenna gain term depends on the adjusted downtilt angle and the user's actual azimuth, reflecting the effect of beam pointing optimization on coverage improvement; the path loss term is based on a standard propagation model, ensuring the universality of the prediction under different environments.

[0046] In one specific implementation, the formula for calculating the user u's experience quality index after the candidate adjustment action is executed is as follows: .

[0047] in, The user experience quality metrics after the candidate adjustment action is executed; For rate weighting coefficients; The predicted downlink rate for user u after the candidate adjustment action is executed; This is the rate normalization parameter; This is the time delay weighting coefficient; This is the predicted service latency value for user u after the candidate adjustment action is executed; This is the time delay normalization parameter.

[0048] In this embodiment, a user experience quality index calculation formula based on downlink rate prediction and service latency prediction is adopted. This formula quantifies the impact of candidate adjustment actions on users' subjective experience into a comparable experience quality index. By introducing rate and latency normalization parameters, the formula can be adapted to different service types (such as video, voice, and web browsing) and network configurations. This allows the optimization process to focus not only on network-side KPIs but also directly on the user's real experience, thereby improving user satisfaction and service perception.

[0049] In one specific implementation, the formula for calculating the total number of handovers by users within cell c after the candidate adjustment action is executed is as follows: .

[0050] in, This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Let be the set of users connected to cell c; 1{·} is an indicator function that takes the value 1 when the condition is true, and 0 otherwise; The RSRP value received by user u from cell c after the candidate adjustment action is executed; The individual offset of cell c after the candidate adjustment action is executed; The RSRP value received by user u from neighboring cell i at the current moment; This represents the individual offset of neighboring cell i at the current moment.

[0051] In this embodiment, a handover frequency prediction formula is adopted by comparing the RSRP of the serving cell and neighboring cells, as well as the individual offset (comparing the maximum value among all neighboring cells for each user). This allows for the early estimation of the handover frequency that may be triggered by candidate adjustment actions. This prediction method aligns with the behavior of users selecting only the optimal neighboring cell during actual handover processes, avoiding duplicate counting and improving the predictive accuracy. The handover penalty weight in the utility function effectively suppresses unnecessary frequent handovers, reducing signaling overhead and network oscillation risk.

[0052] In one specific implementation, the formula for calculating the change in energy consumption of cell c after the candidate adjustment action is executed is as follows: .

[0053] in, This represents the change in energy consumption of cell c after the candidate adjustment action is executed. The energy consumption value of cell c after the candidate adjustment action is executed; This represents the energy consumption value of cell c at the current moment. Preset RF power efficiency factor; The adjusted transmit power of cell c after the candidate adjustment action is executed; The load rate of cell c after the candidate adjustment action is executed; This represents the transmit power of cell c at the current moment; This represents the load rate of cell c at the current moment.

[0054] In this embodiment, a formula for calculating energy consumption changes based on variations in transmit power and load rate is used to quantify the impact of candidate adjustment actions on cell energy consumption. This formula incorporates load rate prediction, reflecting the combined effect of candidate adjustment actions on base station power consumption by adjusting transmit power and changing cell load rate. This allows the optimization process to improve performance and user experience while also considering energy efficiency.

[0055] In one specific embodiment, step S104 specifically includes the following steps S1041 to S1043.

[0056] S1041. Select candidate adjustment actions that meet preset constraints from the set of candidate adjustment actions for the cell, and use them as valid candidate adjustment actions to form a subset of valid candidate adjustment actions; S1042. Select effective candidate adjustment actions that meet the preset amplitude limit conditions from the subset of effective candidate adjustment actions, and use them as feasible adjustment actions to form a subset of feasible adjustment actions; S1043. Select the feasible adjustment action with the largest target utility value from the subset of feasible adjustment actions as the optimal adjustment action.

[0057] In this embodiment, a hierarchical screening method is adopted (first screening valid candidate adjustment actions that meet the constraints, then screening feasible adjustment actions that meet the amplitude limit conditions, and finally selecting the feasible adjustment action with the largest target utility value). This method can eliminate infeasible or unsafe adjustment schemes layer by layer, ensuring that the final adjustment action not only meets basic constraints such as coverage and switching, but also meets the adjustment step size limit, thereby taking into account both optimization effect and system stability, and avoiding increased computational overhead due to invalid search.

[0058] In one specific implementation, the preset constraints include: Coverage constraints: in, For the i-th candidate adjustment action The RSRP value of user u after execution; This is the preset lower limit of signal strength.

[0059] Switching count constraint: .in, For the i-th candidate adjustment action The total number of handovers by users within cell c after execution; This is the preset maximum number of switching attempts.

[0060] Physical parameter limitations: ; ; .

[0061] in, This is the preset lower limit of the cell's transmit power; The cell transmit power corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit of cell transmit power; This is the preset lower limit of the antenna downtilt angle; The antenna downtilt angle corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit of the antenna downtilt angle; This is the preset lower limit for individual offsets within the community; The cell offset corresponding to the optimal adjustment action executed in the previous cycle; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset within the community.

[0062] In this embodiment, by setting coverage constraints, handover count limits, and physical parameter range constraints, it can be ensured that candidate adjustment actions do not generate coverage holes, cause signaling overload, or exceed the physical capabilities of the equipment. These constraints jointly safeguard the safety boundaries of the adjustment actions from three dimensions: coverage quality, signaling load, and equipment security, effectively avoiding the risk of network performance degradation caused by parameter adjustments.

[0063] In one specific implementation, the preset amplitude limiting condition includes: ; ; .

[0064] in, For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit for power adjustment; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit for antenna downtilt adjustment; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset adjustment within the community.

[0065] In this embodiment, by setting an upper limit on the adjustment range of each physical parameter (transmit power, antenna downtilt angle, and individual cell offset), it is possible to avoid drastic fluctuations in network performance or dropped calls due to excessively large single adjustment ranges in candidate adjustment actions. This range limit complements the aforementioned physical parameter range constraints; the former ensures that the state after execution does not exceed the limits, while the latter ensures that the adjustment process itself does not change abruptly.

[0066] The cell parameter adaptive adjustment method provided in this invention solves the problem of traditional methods lacking proactive response and multi-parameter coordination capabilities by repeatedly executing a closed-loop process of candidate adjustment action generation, multi-objective utility function evaluation, constraint screening, and optimal adjustment action execution within each scheduling cycle. This process incorporates physical parameters such as cell transmit power, antenna downtilt angle, and individual cell offset into a collaborative optimization framework. Based on prediction methods such as Shannon's formula, path loss model, antenna directional gain, and queuing theory, it can accurately estimate the changes in throughput, SINR, RSRP, QoE, handover count, and energy consumption after executing candidate adjustment actions. Furthermore, it comprehensively balances multiple objectives such as network performance, user experience, handover overhead, energy consumption, and edge user ratio through a weighted utility function. Simultaneously, it optimizes the adjustment action under constraints such as coverage safety threshold, handover count limit, physical parameter range, and single adjustment magnitude. This enables rapid response to sudden traffic surges and changes in user distribution, avoids the side effects of single-parameter optimization, improves edge user coverage quality, reduces signaling overhead and equipment energy consumption, and ultimately achieves a significant improvement in overall network capacity and operational efficiency.

[0067] Figure 2 This is a flowchart illustrating another cell parameter adaptive adjustment method provided in an embodiment of the present invention. Figure 2 As shown, the method includes the following steps S201 to S205.

[0068] S201. Data Acquisition and Preprocessing.

[0069] The purpose of this step is to accurately and in real-time collect various key performance indicators (KPIs) and user-related information in the network, and to effectively clean and statistically analyze the collected raw data, remove outliers and noise, and form stable input features to provide reliable basic data support for subsequent load prediction, user distribution prediction and decision optimization.

[0070] Because wireless network environments are complex and signal measurements fluctuate, directly using raw data can easily lead to misjudgments, resulting in inaccurate predictions and unstable control. Therefore, it is essential to extract indicators that accurately reflect the true network condition through appropriate statistical and filtering methods.

[0071] This step requires collecting the following data: 1) Base station level indicators, including: The current number of PRBs (Physical Resource Blocks) possessed by cell c is PRB_total. c (t) and the amount of PRB already used (PRB_used) c (t), used to calculate the cell load rate; The number of active users in cell c at the current moment is N user,c (t); The transmit power P of cell c at the current moment t,c (t) and antenna downtilt angle θ c (t), which is subsequently used for physical parameter adjustment; Current cell C handover count statistics HO c (t) (Frequent switching indicates network instability); The average RSRP value reported by users in cell c at the current time and average SINR value This reflects the signal quality.

[0072] 2) User-level metrics, including: Latitude and longitude information for each user; Physical layer signal metrics reported by each user: RSRP (Reference Signal Receiving Power), SINR (Signal to Interference plus Noise Ratio), etc. Real-time downlink rate R u (t) and business latency index D u (t) (used to estimate user experience QoE).

[0073] 3) External features: timestamp (time, day of the week, holiday identifier); event information that may affect traffic (concerts, holiday promotions, etc.).

[0074] These data are periodically reported by base stations, core networks, and user terminals, and collected by a centralized system.

[0075] The raw data collected above is processed as follows: Set the time window length W agg For example, 60 seconds, calculate the statistical characteristics within the sliding window.

[0076] The formula for calculating the community load rate is as follows: L c (t)=PRB_used c (t) / PRB_total c (t).

[0077] Among them, PRB_used c (t) represents the number of PRBs (Physical Resource Blocks) used in cell c at time t; PRB_total c (t) represents the number of PRBs (Physical Resource Blocks) owned by cell c at time t; t is the current time.

[0078] Based on the current and historical load rates of the cell, the average load of the sliding window is calculated using the following formula:

[0079] Calculating the average load using a sliding window can smooth out instantaneous fluctuations and reflect the changing trend of the cell's load.

[0080] User QoE Metric Construction: Define the Quality of Experience (QoE) metric for each user as follows:

[0081] Where: R u (t) is the user's current downlink rate; D u (t) represents the user's current service latency; R max It is the rate normalization parameter (such as the maximum observation or tolerance limit); D max w1 and w2 are latency normalization parameters (such as maximum observation value or tolerance upper limit); w1 and w2 are weighting coefficients that reflect the relative importance of rate and latency to the experience, and satisfy w1 + w2 = 1.

[0082] The average QoE of all users in cell c is calculated to obtain the cell-level average QoE.

[0083] Quantile filtering is used on extreme values ​​in the collected raw data (e.g., removing the highest 1% and lowest 1% of data) to prevent abnormal sampling from interfering with statistics.

[0084] After the aforementioned processing, the current multidimensional network state vector of each cell is output. This vector is stable and accurate, and is used for subsequent prediction and decision-making.

[0085] in, The average load of the cell's sliding window at time t reflects the cell's resource utilization. : Average user experience (QoE) metric for the cell at time t; The average RSRP value of users in the cell at time t reflects the average signal coverage strength of users in the cell. The average SINR value of users within a cell at time t reflects the average signal quality of users within the cell; HO c (t): Number of cell handovers at time t, measuring network stability; N user,c (t): Number of active users in the cell at time t, used to assist in load prediction.

[0086] This step ensures that the system can make predictions and optimizations based on real and stable network operating conditions, fundamentally guaranteeing the scientific nature and effectiveness of dynamic adjustments to cell boundaries.

[0087] S202. Short-term forecasting of load and user distribution.

[0088] The core objective of this step is to predict load changes in each cell within a short future time window, proactively identifying potential overload or resource idleness; predict changes in user spatial distribution, understanding user migration trends in different cells or edge areas; and provide forward-looking information for dynamic adjustment of cell boundaries and resource scheduling, enabling proactive optimization rather than reactive responses. Through this prediction, measures can be taken before network load surges (adjusting transmit power, offset, or resource allocation) to ensure user experience and system stability.

[0089] 1) Short-term load forecast for the community: Load forecasting models can use autoregressive models with exogenous variables (ARX models) to predict future loads.

[0090] The ARX model (Auto-Regressive with Exogenous Inputs) is a classic time series forecasting method. It predicts future values ​​by linearly combining two parts of information: an autoregressive component, which utilizes several historical observations of the target variable (e.g., cell load); and an exogenous input component, which introduces external variables related to the target variable that can be obtained in advance (e.g., time, holiday markers, event indicators, etc.). This model is simple in structure, computationally efficient, and can capture the trend of load evolution over time and the influence of the external environment.

[0091] The formula for predicting the load of a residential community is as follows: .

[0092] in, The load rate of cell c at time (t+τ) is the predicted load rate of cell c in the τth step in the future. Each step is equal to one sampling period (e.g., 60 seconds). The load rate of cell c at time (t+τ-i) is the historical load data of cell c, and is an autoregressive component. : Autoregressive coefficients, reflecting the weight of historical load on future load; : Exogenous variables of cell c at future time (t+τ), including time stamps, holiday markers, event markers, etc., reflecting the impact of the environment on the load; γ: Variable coefficients; : Noise term, representing unpredictable random fluctuations; n: Historical order, usually taken as the data length within the past 5 to 10 minutes, balancing the impact of history and computational complexity.

[0093] Specifically, historical data can be used to train coefficients. For i and γ, the training method usually adopts least squares or recursive least squares (RLS).

[0094] In this step, each cell is modeled independently. Neighbor cell load can also be introduced as a covariate to improve prediction accuracy.

[0095] This step outputs the future short-term load sequence for each cell: ,in ,and This represents the maximum number of prediction steps.

[0096] 2) Short-term prediction of user distribution: For the distribution of users in a cell, first establish the state transition probability matrix P, and then predict the user distribution vector at the next time step based on the user distribution vector at the previous time step and the state transition matrix P: .

[0097] in, : The user distribution vector of cell c at time (t+τ) can divide the cell coverage area into multiple sub-regions (such as grids), and the user distribution vector can be the proportion of users in each sub-region; P: The state transition probability matrix of the sub-regions within the cell, consisting of elements p ij Composition, p ij This represents the probability that a user migrates from sub-region i to sub-region j, obtained based on historical trajectory statistics; : The user distribution vector of cell c at time (t+τ-i).

[0098] This step outputs the future short-term user distribution sequence for each cell: ,in ,and This represents the maximum number of prediction steps.

[0099] S203. Candidate adjustment actions are generated.

[0100] The core objective of this step is to generate a set of executable cell physical parameter adjustment schemes, while ensuring network security and stability, providing a set of optional adjustment actions for utility function evaluation and optimal decision-making; to evaluate the impact of each adjustment action on load, user experience, and handover behavior through simulation prediction results; and to avoid directly optimizing continuous parameters significantly, thereby reducing computational complexity and ensuring real-time executability. Therefore, this step transforms the predicted future state into a set of candidate adjustment actions for specific and controllable operations, providing alternative solutions for subsequent optimization selection.

[0101] This step requires the following data to be entered: The transmit power P of cell c at the current moment t,c ; The antenna downtilt angle θ of cell c at the current moment c ; Individual bias in cell c c (Cell Individual Offset, CIO); Future time window cell load prediction L c (t+τ), Prediction of User Distribution in the Cellular Area D c (t+τ) (obtained from step S202).

[0102] The adjustment range restrictions are as follows: Cell C transmit power allowable adjustment range ΔP t,c ∈[P min ,P max ], step size 1~3 dB; The allowable adjustment range of the downtilt angle of the cell's C antenna is Δθ. c ∈[θ min ,θ max Step size 1°~3°; Allowable adjustment range Δbias for individual offset in cell c c ∈[bias min bias max ], step size 3~6 dB.

[0103] The constraints are as follows: Minimum Signal Response Threshold (SINR) for Cell Coverage min Maximum number of cell handovers (HO) max Physical limitations on cell transmit power and cell antenna angle range.

[0104] The generation of candidate adjustment actions involves the following steps: 1. Define the parameter adjustment range: Based on the predicted vector of user distribution in the cell, feasible discrete adjustment values ​​are set for the physical parameters of each cell.

[0105] Specifically, the coverage area of ​​cell c is divided into K sub-regions (grids), and the user distribution prediction vector output in step S202 is D. c (t+τ) = [d1, d2, …, d K ], where d j This represents the proportion of users located in sub-region j at the future time (t+τ). Based on the predicted user proportions in each sub-region, the discrete adjustment step size and value range of the three physical parameters in the candidate actions are dynamically adjusted.

[0106] (1) Transmit power adjustment step size ΔP t,c Adaptive determination.

[0107] The set S of marginal sub-regions in the statistical user distribution prediction vector edge User ratio: .

[0108] If ρ edge If the percentage exceeds a preset threshold (e.g., 30%), it indicates a future concentration of edge users, necessitating enhanced coverage; therefore, ΔP should be increased. t,c The positive adjustment step size (e.g., expanding from the default {0, ±1 dB} to {0, ±1, ±2 dB}) allows for a greater power boost; conversely, if ρ edge If the user base is low and concentrated near the center, a smaller step size is used to save energy.

[0109] (2) Antenna downtilt adjustment step size Δθ c Adaptive determination.

[0110] Calculate the centroid sub-region (i.e., the sub-region with the largest user proportion) in the user distribution prediction vector and its deviation angle relative to the current antenna main lobe direction. If the centroid offset exceeds a preset threshold (e.g., 5°), increase Δθ. c The step size (e.g., expanding from {0, ±1°} to {0, ±2°, ±3°}) is used to quickly adjust the coverage direction; if the center of gravity is consistent with the current antenna main lobe direction, a smaller step size is used for fine-tuning.

[0111] (3) Individual bias in the community Δbias c Adaptive determination.

[0112] Analyze the changing trend of the user proportion in the sub-region at the boundary between the serving cell and neighboring cells in the user distribution prediction vector. If the prediction shows a significant increase in the user proportion in the sub-region at the boundary (indicating an increase in potential handover demand), then increase Δbias. c The step size can be increased (e.g., from {0, ±1 dB} to {0, ±2, ±3 dB}) to quickly respond to load balancing needs; conversely, a smaller step size can be used to maintain stability.

[0113] Specific examples are as follows:

[0114] These distances ensure that the granularity of adjustment actions is small enough to avoid network oscillations, while also controlling the total amount of adjustment actions to facilitate real-time calculation; the adjustment range is subject to physical limitations and interference constraints to ensure safe execution.

[0115] 2. Enumerate candidate adjustment action combinations: For each cell, the discrete values ​​of the three parameters—cell transmit power, antenna downtilt angle, and individual cell offset—are multiplied by a Cartesian product to form a candidate adjustment action set C. c : C c ={(P t,c +ΔP t,c ), (θ c +Δθ c (bias) c +Δbias c )}.

[0116] The Cartesian product refers to listing all possible combinations of elements from multiple sets to form a complete set consisting of ordered pairs or ordered tuples. The purpose of performing a Cartesian product on the discrete values ​​of these three parameters is to exhaustively enumerate all possible parameter adjustments and combinations.

[0117] For each cell, the cell transmit power P must be considered simultaneously. t,c , antenna downtilt angle θ c Individual bias in the community c These three parameters can be adjusted in different ways. First, assign a set of discrete adjustment values ​​to each parameter (e.g., transmit power can be adjusted by ±2dB, antenna angle by ±5°, and individual offset by ±3dB). Then, list all possible combinations of the three parameters, i.e., enumerate all possible adjustments, and generate a set of candidate adjustment actions C through a Cartesian product. c This serves as a pool of potential solutions.

[0118] In the current parameter (P) t,c , θ c bias cBased on this, add the possible adjustment amounts (ΔP) respectively. t,c , Δθ c ,Δbias c This forms a new set of parameter combinations. Listing all possible combinations yields all candidate adjustment actions for cell c, forming the candidate adjustment action set C. c The subsequent algorithm will select the optimal adjustment action from these candidate adjustments and execute it.

[0119] 3. For each candidate adjustment action, estimate its impact on cell network metrics after execution: 1) Predict RSRP changes.

[0120] The path loss between user u and the base station is derived using the standard path loss model:

[0121] Among them, PL(d u ):d u The path loss (dB) at point d u PL0: Path loss (dB) at reference distance d0, typically taken as 1 meter (indoors); n: Environmental attenuation index (e.g., 2.8~3.5 in urban areas); X σ Log-normal shading fading (dB) follows an overall distribution with a mean of 0 and a standard deviation of σ, reflecting random fluctuations caused by environmental shading.

[0122] Adjusted transmit power With path loss Perform calculations to predict the adjusted RSRP:

[0123] in, : The RSRP value (predicted value) of user u after the candidate adjustment action is executed; : The adjusted transmit power of cell c after the candidate adjustment action is executed.

[0124] If the candidate adjustment actions include antenna downtilt adjustment, directional gain must also be considered:

[0125] in, After the candidate adjustment action is executed, the antenna gain of cell c is adjusted, specifically the antenna downtilt angle of cell c is adjusted. Below, the antenna gain in the user's u direction.

[0126] 2) Predicting SINR changes:

[0127] in, : The SINR value (predicted value) of user u after the candidate adjustment action is executed; After the candidate adjustment action is executed, the RSRP value (predicted value) of user u is equivalent to the service signal power received by user u from the serving cell after the candidate adjustment action is executed. : The interference power (measured value) received by user u from neighboring cell i at the current moment; : The sum of interference power received by user u from all neighboring cells of the serving cell at the current moment; N0: Noise power (system constant).

[0128] 3) Throughput estimation.

[0129] The throughput of user u can be approximated using the Shannon formula. Specifically, using the Shannon channel capacity formula, the data transmission rate that the user may achieve after cell transmit power adjustment is estimated based on the predicted SINR. The specific calculation formula is as follows:

[0130] in, : The throughput (predicted value) of user u after the candidate adjustment action is executed; The bandwidth allocated to user u is based on known system parameters (such as bandwidth per PRB × number of allocated PRBs), and the subsequent scheduling module will further allocate resources. SINR value (predicted value) for user u after the candidate adjustment action is executed.

[0131] 4) QoE estimation.

[0132]

[0133] in, : The QoE metric (predicted value) of user u after the candidate adjustment action is executed; Rate weighting coefficient; : The predicted downlink rate of user u after the candidate adjustment action is executed; Rate normalization parameter; Delay weighting coefficient; : The predicted service latency for user u after the candidate adjustment action is executed; : Delay normalization parameter.

[0134] After the candidate adjustment action is executed, the formula for calculating the predicted downlink rate of user u is as follows:

[0135] in, Link efficiency factor.

[0136] After the candidate adjustment action is executed, the formula for calculating the predicted service latency for user u is as follows:

[0137] in, : The length of the data queue to be transmitted for user u at the current moment.

[0138] 5) Estimation of the number of handovers.

[0139] Based on the predicted RSRP and individual cell offset changes, the total number of user handovers that may occur in the future is estimated.

[0140]

[0141] in, : The total number of handovers for users in cell c after the candidate adjustment action is executed (predicted value); : The set of users connected to cell c, where cell c is the currently serving cell; 1{·}: An indicator function that takes the value 1 when the condition is true, and 0 otherwise; : The RSRP value (predicted value) received by user u from cell c after the candidate adjustment action is executed; After the candidate adjustment action is executed, the individual cell offset of cell c after adjustment; The RSRP value (measured value) received by user u from neighboring cell i at the current moment, where neighboring cell i is a neighboring cell of the current serving cell; : The individual offset of neighboring cell i at the current moment.

[0142] For each user u, a conditional judgment is performed. If, after the candidate adjustment action is executed, the sum of the RSRP value of the current serving cell and the individual offset received by the user is lower than the sum of the RSRP values ​​and individual offsets received by the user from neighboring cells, then it can be predicted that this user will undergo a handover. This judgment is based on the indicator function 1{ This is implemented by using a checklist, where a value of 1 is recorded when the condition is true, and 0 is recorded otherwise. Finally, the results for all users are summed to obtain the predicted total number of switches (HO). pred .

[0143] 6) Estimation of changes in energy consumption in the community.

[0144]

[0145] in, : The change in energy consumption (predicted value) of cell c after the candidate adjustment action is executed. Energy consumption (predicted value) of cell c after the candidate adjustment action is executed. : The energy consumption value (measured value) of cell c at the current moment; RF power efficiency factor (0.5~1.0), which is related to power amplifier design, frequency band, etc. : The adjusted transmit power (adjustment value) of cell c after the candidate adjustment action is executed. After the candidate adjustment action is executed, the load rate (predicted value) of cell c can be obtained by using the future cell load rate L in step S202. c (t+τ) is used as an approximation because, within the short-term forecast window, the total user demand is relatively stable. Candidate adjustment actions primarily redistribute load rather than change the total load, and the load forecast in step S202 is modeled based on historical trends and external events (such as holidays). Therefore, without introducing additional complex models, the load rate predicted in step S202 can be used as a reasonable approximation of the load rate after the candidate adjustment actions are executed. For a more accurate estimate, fine-tuning can be done by combining the number of switching users estimated in step S203 (e.g., adjusting based on the difference in the number of users switching in and out using the average load per user). : The current transmit power (measured value) of cell c; : The load rate (measured value) of cell c at the current moment.

[0146] The output of this step is the set of candidate adjustment actions C for cell c. c .

[0147] S204. Design and Constraints of Objective Utility Functions.

[0148] The core objective of this step is to quantify the comprehensive impact of each candidate adjustment action in the candidate adjustment action set on network performance, user experience, switching costs, and energy consumption; to provide clear evaluation criteria for the optimization module, enabling the system to automatically select the optimal adjustment action; and to ensure that the selection of candidate adjustment actions takes into account multiple objectives, avoiding side effects caused by optimizing a single indicator (such as increased throughput but decreased user edge experience or increased switching frequency). The predicted effects of candidate adjustment actions are converted into comparable comprehensive utility values ​​as the basis for optimization decisions.

[0149] The objective utility function of community c adopts a multi-objective weighted form to quantify candidate adjustment actions:

[0150] in, Throughput weight; User experience weight; Switch penalty weights; : Energy consumption penalty weight; log(·): Logarithmic function, whose base can be any constant greater than 1, such as 2, 10 or e (natural logarithm), because the base of the logarithm does not affect monotonicity and relative comparison; Zero-prevention coefficient, used to prevent the logarithm from being zero; : The throughput (predicted value) of user u after the candidate adjustment action is executed; : The QoE metric (predicted value) of user u after the candidate adjustment action is executed; : The total number of handovers for users in cell c after the candidate adjustment action is executed (predicted value); : The change in energy consumption (predicted value) of cell c after the candidate adjustment action is executed.

[0151] After the candidate adjustment action is executed, the following constraints must be met to ensure network security and coverage: 1) Coverage constraint: RSRP u ≥RSRP min , u∈Service area, the service area refers to the coverage area of ​​the serving cell (cell c), ensuring that all users accessing the serving cell after adjustment receive sufficient signal strength; 2) Switching count constraint HO pred (c)≤HO max To avoid frequent switching caused by adjustment actions and reduce signaling load; 3) Physical parameter limitations: P t,min ≤(P t,c +ΔP t,c )≤P t,max ; θ min ≤(θ c +Δθ c )≤θ max ; bias min ≤(bias c +Δbias c )≤bias max .

[0152] Calculate the set of candidate adjustment actions C c The target utility value U after each candidate adjustment action is executed is determined, and it is checked whether it meets the constraints. All candidate adjustment actions that meet the constraints constitute a subset C of valid candidate adjustment actions. valid ={ a 1, a 2,..., a N}

[0153] For C valid Each valid candidate adjustment action a i For each element i = 1, 2, ..., N, the following constraints must be satisfied: ; ; ; ; .

[0154] in, It is the parameter setting from the last time (i.e., the parameter value set by the adjustment action executed in the previous cycle).

[0155] Output of this step: Subset C of valid candidate adjustment actions valid and C valid The target utility value U corresponding to each valid candidate adjustment action.

[0156] S205. Optimize selection and execution decisions.

[0157] The purpose of this step is to select the optimal adjustment action from the subset of effective candidate adjustment actions to maximize the utility function while satisfying the constraints; to control the execution frequency and magnitude of adjustment actions to avoid network oscillations or user experience degradation caused by frequent adjustments; to realize dynamic cell boundary adjustment and resource optimization to ensure stable operation of the system in complex environments; and to provide a real-time execution plan to ensure that adjustment actions can be implemented quickly and reliably, thereby transforming predicted and candidate adjustment actions into actual executable operations.

[0158] This step requires the following input information: Valid candidate adjustment action subset C valid ={ a 1, a 2,..., a N}

[0159] The target utility value U corresponding to the effective candidate adjustment action. a i); Community historical adjustment information: Execution time of the last adjustment action t last ; Adjust the amplitude limit (to avoid frequent adjustments or excessive parameter changes, check whether the valid candidate adjustment actions meet the maximum amplitude limit): ; ; ; .

[0160] Adjust the action subset C from valid candidates validFurther filtering reveals adjustment actions that meet the limitations of the adjustment range, resulting in a feasible subset C of adjustment actions. feasible .

[0161] Optimal adjustment action selection: From the subset C of executable adjustment actions feasible Choose the adjustment action with the highest utility value As the optimal adjustment action:

[0162] If multiple adjustment actions correspond to the same maximum utility value, the minimum adjustment principle can be adopted, selecting the adjustment action that is closest to the last parameter setting as the optimal adjustment action to reduce network oscillations.

[0163] Execution of adjustment actions (sending the parameters of the optimal adjustment actions to the base station control unit): ; ; .

[0164] After the optimal adjustment action is executed, the execution time and parameter settings of this adjustment action are recorded to provide a reference for the parameter adjustment in the next cycle.

[0165] Output of this step: The parameters for the optimal adjustment action to be executed in the current scheduling cycle are: cell transmit power, antenna downtilt angle, and individual offset adjustment value. Adjust the action execution status: success / failure indicator for easy logging and exception handling; Updated community status: Provides input data for the next round of indicator collection.

[0166] After this step is completed, the system can achieve closed-loop optimization of the entire process.

[0167] In summary, step S201 provides reliable input, step S202 predicts the future network state, step S203 generates executable candidate adjustment actions, step S204 quantifies the effect of candidate adjustment actions, and step S205 selects the optimal candidate adjustment action and executes it.

[0168] The cell parameter adaptive adjustment method provided in this invention collects multi-dimensional data such as user location, signal quality, and antenna parameters, and uses a prediction model to calculate short-term predictions of load and user distribution over a future period. Based on this, candidate adjustment action combinations for multiple parameters such as cell transmit power, antenna downtilt angle, and individual cell offset are generated. An optimization algorithm selects the optimal adjustment scheme and implements it in the network. Simultaneously, the prediction model is adaptively corrected based on the results of the adjustment actions, thereby achieving dynamic optimization of cell parameters and improving network performance.

[0169] Figure 3 This is a schematic diagram of the cell parameter adaptive adjustment device provided in an embodiment of the present invention. Figure 3 As shown, the device includes the following modules that work together in each scheduling cycle: a candidate adjustment action generation module 301, a target utility function construction module 302, a target utility value calculation module 303, an optimal adjustment action determination module 304, and an optimal adjustment action execution module 305.

[0170] The candidate adjustment action generation module 301 is configured to set discrete feasible adjustment values ​​based on cell physical parameters, and superimpose the current cell physical parameters with the corresponding feasible adjustment values ​​to obtain the adjustment action for each physical parameter. Then, the adjustment actions of various physical parameters are combined to generate a set of cell candidate adjustment actions. The physical parameters include at least one of cell transmit power, antenna downtilt angle, and cell individual offset. The target utility function construction module 302 is configured to construct a cell target utility function based on multiple target parameters; wherein, the target parameters involve at least two of network performance, user experience, handover overhead, and energy consumption; The target utility value calculation module 303 is configured to calculate the target utility value after the execution of each candidate adjustment action in the cell candidate adjustment action set based on the cell target utility function; The optimal adjustment action determination module 304 is configured to select the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions of the cell, under the premise of satisfying preset constraints, as the optimal adjustment action; The optimal adjustment action execution module 305 is configured to execute the optimal adjustment action in the current scheduling period.

[0171] In one specific implementation, the target utility function constructed by the target utility function construction module 302 is as follows: .

[0172] in, Let c be the target utility function; Let be the set of users connected to community c; Preset throughput weights; The throughput of user u after the candidate adjustment action is executed; To prevent a zero coefficient; Preset user experience weights; The user experience quality metrics after the candidate adjustment action is executed; Preset switching penalty weights; This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Preset energy consumption penalty weights; This represents the change in energy consumption of cell c after the candidate adjustment action is executed.

[0173] In one specific implementation, the formula for calculating the throughput of user u after the candidate adjustment action is executed is as follows: .

[0174] in, The throughput of user u after the candidate adjustment action is executed; The bandwidth allocated to user u; The SINR value of the user after the candidate adjustment action is executed; .

[0175] in, The SINR value of the user after the candidate adjustment action is executed; The RSRP value of the user after the candidate adjustment action is executed; This represents the interference power received by user u from neighboring cell i at the current moment. This is the sum of the interference power received by user u from all neighboring cells of cell c at the current moment; Preset noise power; .

[0176] in, The RSRP value of user u after the candidate adjustment action is executed; The adjusted transmit power of cell c after the candidate adjustment action is executed; for Path loss at that location, and The distance from user u to the base station; This refers to the antenna gain in the user's u direction at the adjusted antenna downtilt angle in cell c after the candidate adjustment action is executed.

[0177] In one specific implementation, the formula for calculating the user u's experience quality index after the candidate adjustment action is executed is as follows: .

[0178] in, The user experience quality metrics after the candidate adjustment action is executed; For rate weighting coefficients; The predicted downlink rate for user u after the candidate adjustment action is executed; This is the rate normalization parameter; This is the time delay weighting coefficient; This is the predicted service latency value for user u after the candidate adjustment action is executed; This is the time delay normalization parameter.

[0179] In one specific implementation, the formula for calculating the total number of handovers by users within cell c after the candidate adjustment action is executed is as follows: .

[0180] in, This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Let be the set of users connected to cell c; 1{·} is an indicator function that takes the value 1 when the condition is true, and 0 otherwise; The RSRP value received by user u from cell c after the candidate adjustment action is executed; The individual offset of cell c after the candidate adjustment action is executed; The RSRP value received by user u from neighboring cell i at the current moment; This represents the individual offset of neighboring cell i at the current moment.

[0181] In one specific implementation, the formula for calculating the change in energy consumption of cell c after the candidate adjustment action is executed is as follows: .

[0182] in, This represents the change in energy consumption of cell c after the candidate adjustment action is executed. The energy consumption value of cell c after the candidate adjustment action is executed; This represents the energy consumption value of cell c at the current moment. Preset RF power efficiency factor; The adjusted transmit power of cell c after the candidate adjustment action is executed; The load rate of cell c after the candidate adjustment action is executed; This represents the transmit power of cell c at the current moment; This represents the load rate of cell c at the current moment.

[0183] In one specific implementation, the step of selecting the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions for the cell, under the premise of satisfying preset constraints, as the optimal adjustment action, includes: From the set of candidate adjustment actions for the cell, candidate adjustment actions that meet the preset constraints are selected as valid candidate adjustment actions, and constitute a subset of valid candidate adjustment actions; Select valid candidate adjustment actions that meet the preset amplitude limit conditions from the subset of valid candidate adjustment actions, and use them as feasible adjustment actions to form a subset of feasible adjustment actions; Select the feasible adjustment action with the largest target utility value from the subset of feasible adjustment actions as the optimal adjustment action.

[0184] In one specific implementation, the preset constraints include: Coverage constraints: .

[0185] in, For the i-th candidate adjustment action The RSRP value of user u after execution; This is the preset lower limit of signal strength.

[0186] Switching count constraint: .

[0187] in, For the i-th candidate adjustment action The total number of handovers by users within cell c after execution; This is the preset maximum number of switching attempts.

[0188] Physical parameter limitations: ; ; .

[0189] in, This is the preset lower limit of the cell's transmit power; The cell transmit power corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit of cell transmit power; This is the preset lower limit of the antenna downtilt angle; The antenna downtilt angle corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit of the antenna downtilt angle; This is the preset lower limit for individual offsets within the community; The cell offset corresponding to the optimal adjustment action executed in the previous cycle; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset within the community.

[0190] In one specific implementation, the preset amplitude limiting condition includes: ; ; .

[0191] in, For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit for power adjustment; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit for antenna downtilt adjustment; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset adjustment within the community.

[0192] The cell parameter adaptive adjustment device provided in this invention solves the problem of traditional methods lacking proactive response and multi-parameter coordination capabilities by repeatedly executing a closed-loop process of candidate adjustment action generation, multi-objective utility function evaluation, constraint screening, and optimal adjustment action execution in each scheduling cycle. This process incorporates physical parameters such as cell transmit power, antenna downtilt angle, and individual cell offset into a collaborative optimization framework. Based on prediction methods such as Shannon's formula, path loss model, antenna directional gain, and queuing theory, it can make relatively accurate forward-looking estimates of throughput, SINR, RSRP, QoE, handover count, and energy consumption changes after executing candidate adjustment actions. Furthermore, it comprehensively balances multiple objectives such as network performance, user experience, handover overhead, energy consumption, and edge user ratio through a weighted utility function. Simultaneously, it optimizes the adjustment action under constraints such as coverage safety threshold, handover count limit, physical parameter range, and single adjustment magnitude. This enables rapid response to sudden traffic and user distribution changes, avoids the side effects of single-parameter optimization, improves edge user coverage quality, reduces signaling overhead and equipment energy consumption, and ultimately achieves a significant improvement in overall network capacity and operational efficiency.

[0193] Based on the same technical concept, embodiments of the present invention also provide a computer device, such as... Figure 4 As shown, the computer device includes a memory 401 and a processor 402. The memory 401 stores a computer program. When the processor 402 runs the computer program stored in the memory 401, the processor 402 executes the aforementioned cell parameter adaptive adjustment method.

[0194] Based on the same technical concept, embodiments of the present invention also provide a computer-readable storage medium storing a computer program thereon, wherein when the computer program is executed by a processor, the processor executes the aforementioned cell parameter adaptive adjustment method.

[0195] In summary, the cell parameter adaptive adjustment method, apparatus, computer equipment, and storage medium provided in this embodiment of the invention introduce a closed-loop optimization mechanism that combines user mobility trend prediction with dynamic network status assessment. Unlike traditional parameter adjustment methods that rely on historical statistics or fixed thresholds, this method couples the prediction results with real-time network resource conditions to generate a multi-dimensional parameter collaborative optimization scheme that includes transmit power, antenna downtilt angle, and individual cell offset. Furthermore, it adaptively corrects the prediction model through execution feedback, thereby achieving dynamic, accurate, and forward-looking cell parameter adjustment.

[0196] It will be understood by those skilled in the art that all or some of the steps, systems, or apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0197] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for adaptive adjustment of cell parameters, characterized in that, This includes the following steps, which are repeated within each scheduling cycle: Discrete feasible adjustment values ​​are set based on the cell physical parameters, and the current cell physical parameters are superimposed with the corresponding feasible adjustment values ​​to obtain the adjustment action for each physical parameter. Then, the adjustment actions of various physical parameters are combined to generate a set of candidate cell adjustment actions. The physical parameters include at least one of cell transmit power, antenna downtilt angle and cell individual offset. A target utility function for a cell is constructed based on multiple target parameters; wherein, the target parameters involve at least two of network performance, user experience, handover overhead, and energy consumption. Based on the cell target utility function, calculate the target utility value after the execution of each candidate adjustment action in the cell candidate adjustment action set; Under the premise of satisfying preset constraints, the candidate adjustment action with the largest target utility value is selected from the set of candidate adjustment actions for the cell as the optimal adjustment action; The optimal adjustment action is executed in the current scheduling cycle.

2. The cell parameter adaptive adjustment method according to claim 1, characterized in that, The target utility function of the cell is as follows: ； in, Let c be the target utility function; Let be the set of users connected to community c; Preset throughput weights; The throughput of user u after the candidate adjustment action is executed; To prevent a zero coefficient; Preset user experience weights; The user experience quality metrics after the candidate adjustment action is executed; Preset switching penalty weights; This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Preset energy consumption penalty weights; This represents the change in energy consumption of cell c after the candidate adjustment action is executed.

3. The cell parameter adaptive adjustment method according to claim 2, characterized in that, The formula for calculating the throughput of user u after the candidate adjustment action is executed is as follows: ； in, The throughput of user u after the candidate adjustment action is executed; The bandwidth allocated to user u; The SINR value of the user after the candidate adjustment action is executed; ； in, The SINR value of the user after the candidate adjustment action is executed; The RSRP value of the user after the candidate adjustment action is executed; This represents the interference power received by user u from neighboring cell i at the current moment. This is the sum of the interference power received by user u from all neighboring cells of cell c at the current moment; Preset noise power; ； in, The RSRP value of user u after the candidate adjustment action is executed; The adjusted transmit power of cell c after the candidate adjustment action is executed; for Path loss at that location, and The distance from user u to the base station; This refers to the antenna gain in the user's u direction at the adjusted antenna downtilt angle in cell c after the candidate adjustment action is executed.

4. The cell parameter adaptive adjustment method according to claim 2, characterized in that, The formula for calculating the user experience quality index after the candidate adjustment action is executed is as follows: ； in, The user experience quality metrics after the candidate adjustment action is executed; For rate weighting coefficients; The predicted downlink rate for user u after the candidate adjustment action is executed; This is the rate normalization parameter; This is the time delay weighting coefficient; This is the predicted service latency value for user u after the candidate adjustment action is executed; This is the time delay normalization parameter.

5. The cell parameter adaptive adjustment method according to claim 2 or 3, characterized in that, The formula for calculating the total number of handovers by users in cell c after the candidate adjustment action is executed is as follows: ； in, This represents the total number of handovers that occurred within cell c after the candidate adjustment action was executed. Let be the set of users connected to cell c; 1{·} is an indicator function that takes the value 1 when the condition is true, and 0 otherwise; The RSRP value received by user u from cell c after the candidate adjustment action is executed; The individual offset of cell c after the candidate adjustment action is executed; The RSRP value received by user u from neighboring cell i at the current moment; This represents the individual offset of neighboring cell i at the current moment.

6. The cell parameter adaptive adjustment method according to claim 2, characterized in that, The formula for calculating the change in energy consumption of cell c after the candidate adjustment action is executed is as follows: ； in, This represents the change in energy consumption of cell c after the candidate adjustment action is executed. The energy consumption value of cell c after the candidate adjustment action is executed; This represents the energy consumption value of cell c at the current moment. Preset RF power efficiency factor; The adjusted transmit power of cell c after the candidate adjustment action is executed; The load rate of cell c after the candidate adjustment action is executed; This represents the transmit power of cell c at the current moment; This represents the load rate of cell c at the current moment.

7. The cell parameter adaptive adjustment method according to claim 1, characterized in that, The step of selecting the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions for the cell, under the premise of satisfying preset constraints, as the optimal adjustment action, includes: From the set of candidate adjustment actions for the cell, candidate adjustment actions that meet the preset constraints are selected as valid candidate adjustment actions, and constitute a subset of valid candidate adjustment actions; Select valid candidate adjustment actions that meet the preset amplitude limit conditions from the subset of valid candidate adjustment actions, and use them as feasible adjustment actions to form a subset of feasible adjustment actions; Select the feasible adjustment action with the largest target utility value from the subset of feasible adjustment actions as the optimal adjustment action.

8. The cell parameter adaptive adjustment method according to claim 7, characterized in that, The preset constraints include: Coverage constraints: ;in, For the i-th candidate adjustment action The RSRP value of user u after execution; The preset lower limit of signal strength; Switching count constraint: ;in, For the i-th candidate adjustment action The total number of handovers by users within cell c after execution; This is the preset maximum number of switching attempts; Physical parameter limitations: ； ； ； in, This is the preset lower limit of the cell's transmit power; The cell transmit power corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit of cell transmit power; This is the preset lower limit of the antenna downtilt angle; The antenna downtilt angle corresponding to the optimal adjustment action performed in the previous cycle; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit of the antenna downtilt angle; This is a preset lower limit for individual offsets within the community; The cell offset corresponding to the optimal adjustment action executed in the previous cycle; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset within the community.

9. The cell parameter adaptive adjustment method according to claim 7, characterized in that, The preset amplitude limit conditions include: ； ； ； in, For the i-th candidate adjustment action The corresponding cell transmit power adjustment amount; This is the preset upper limit for power adjustment; For the i-th candidate adjustment action The corresponding antenna downtilt adjustment amount; This is the preset upper limit for antenna downtilt adjustment; For the i-th candidate adjustment action The corresponding individual offset adjustment amount for the community; This is the preset upper limit for individual offset adjustment within the community.

10. A cell parameter adaptive adjustment device, characterized in that, The following modules work together in each scheduling cycle: The candidate adjustment action generation module is configured to set discrete feasible adjustment values ​​based on cell physical parameters, and superimpose the current cell physical parameters with the corresponding feasible adjustment values ​​to obtain the adjustment action for each physical parameter. Then, the adjustment actions of various physical parameters are combined to generate a set of candidate cell adjustment actions. The physical parameters include at least one of cell transmit power, antenna downtilt angle, and cell individual offset. The target utility function construction module is configured to construct a cell target utility function based on multiple target parameters; wherein, the target parameters involve at least two of network performance, user experience, handover overhead, and energy consumption; The target utility value calculation module is configured to calculate the target utility value after the execution of each candidate adjustment action in the set of candidate adjustment actions for the cell, based on the target utility function of the cell. The optimal adjustment action determination module is configured to select the candidate adjustment action with the largest target utility value from the set of candidate adjustment actions for the cell, under the premise of satisfying preset constraints, as the optimal adjustment action; The optimal adjustment action execution module is configured to execute the optimal adjustment action in the current scheduling period.

11. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the cell parameter adaptive adjustment method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the processor performs the cell parameter adaptive adjustment method according to any one of claims 1 to 9.

Images