Cell network optimization method and device, equipment, storage medium and product

Abstract

The invention relates to a cell network optimization method and device, equipment, a storage medium and a product, and relates to the technical field of communication. The method comprises the following steps: acquiring a performance index matrix of a target cell cluster; the target energy-saving strategy is simulated and executed based on the performance index matrix of the target cell cluster to obtain a simulation execution result, the target energy-saving strategy is an energy-saving strategy in the strategy iterative optimization process, and the energy-saving strategy comprises at least one of the following: changing the activation state of each cell and changing the activation state of each cell; adjusting a target parameter, used for adjusting the coverage range of the cell, of the cell under the condition that the user perception condition is met; evaluating the simulation execution result based on the comprehensive optimization target to obtain an evaluation result; the comprehensive optimization target is constructed based on a user perception index and an energy-saving index; if the evaluation result reaches an optimization completion condition, executing a target energy-saving strategy; through the method, the perception experience of the user is ensured while the energy is saved, and the energy conservation of the base station and the effective optimization of the cell load are realized.

Description

Technical Field

[0001] Embodiments of the present application relate to the field of communication technologies, and particularly to a method, device, equipment, storage medium and product for optimizing a cell network. Background Art

[0002] In the current era of rapid development of the communication industry, as the core infrastructure of the mobile communication network, the number of base stations is increasing continuously, and its energy consumption problem is becoming increasingly prominent.

[0003] Existing base station energy-saving solutions mainly rely on traditional fuzzy logic systems or on historical data and prediction models, lacking the ability to optimize decisions in complex environments and also lacking the ability to dynamically adjust according to real-time network status and user needs. Summary of the Invention

[0004] Embodiments of the present application provide a method, device, equipment, storage medium and product for optimizing a cell network, which can ensure the user perception experience while saving energy, and realize the effective optimization of base station energy saving and cell load. The technical solution is as follows.

[0005] On the one hand, a method for optimizing a cell network is provided, and the method includes:

[0006] Obtain a performance index matrix of a target cell cluster, where each vector in the performance index matrix is used to indicate the performance index of each in-cluster cell in the target cell cluster;

[0007] Based on the performance index matrix of the target cell cluster, simulate the execution of a target energy-saving strategy to obtain a simulation execution result of the target energy-saving strategy. The target energy-saving strategy is an energy-saving strategy in the process of policy iteration optimization, and the energy-saving strategy includes at least one of the following: changing the activation state of each in-cluster cell, and adjusting the target parameters of the in-cluster cell when the user perception condition is met; the target parameters are used to adjust the coverage range of the in-cluster cell;

[0008] Evaluate the simulation execution result of the target energy-saving strategy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving strategy; the comprehensive optimization target is constructed based on user perception indicators and energy-saving indicators;

[0009] When the evaluation result of the target energy-saving strategy reaches the optimization completion condition, execute the target energy-saving strategy.

[0010] In a possible implementation manner, the evaluating the simulation execution result of the target energy-saving strategy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving strategy includes:

[0011] When the user perception metric determined based on the simulation execution result meets the user perception condition, calculate the reward value of the target energy-saving policy through a reward function; the reward function is used to evaluate the energy-saving metric;

[0012] When the user perception metric determined based on the simulation execution result does not meet the user perception condition, set the reward value of the target energy-saving policy to a target value.

[0013] In a possible implementation, the user perception metric includes at least one of the following: residence ratio metric, reference signal received power metric, transmission rate metric, and interference value metric.

[0014] In a possible implementation, when the evaluation result of the target energy-saving policy does not meet the optimization condition, perform policy optimization according to the reward value of the target energy-saving policy until the evaluation result meets the optimization completion condition.

[0015] In a possible implementation, the performance metrics of each cell in a cluster include the number of cell-edge users and the frequency band co-coverage situation; the method further includes:

[0016] Calculate the weak coverage degree of each cell in the target cell cluster respectively to obtain the weak coverage degree of each cell in the cluster;

[0017] Based on the weak coverage degree of each cell in the cluster, determine the number of cell-edge users of each cell in the cluster, and the number of cell-edge users is used to indicate the potential number of edge users of the neighboring cells corresponding to the cell in the cluster;

[0018] Perform frequency band co-coverage analysis on each cell in the cluster to obtain the frequency band co-coverage situation of each cell in the cluster.

[0019] In a possible implementation, when the target cell in the target cluster has edge users, has neighboring cells with frequency band co-coverage, and meets the user perception condition, the target energy-saving policy instructs to adjust the target parameters of the target cell in the target cluster and the neighboring cells of the target cell in the target cluster; the target cell in the target cluster is any cell in the target cell cluster;

[0020] When the target cell in the target cluster does not have neighboring cells with frequency band co-coverage, the target energy-saving policy instructs to skip the parameter adjustment operation of the target cell in the target cluster.

[0021] On the other hand, a cell network optimization device is provided, and the device includes:

[0022] An acquisition module, configured to acquire a performance metric matrix of a target cell cluster, where each vector in the performance metric matrix is used to indicate the performance metrics of each intra-cluster cell in the target cell cluster;

[0023] A policy optimization module, configured to simulate the execution of a target energy-saving policy based on the performance metric matrix of the target cell cluster to obtain a simulation execution result of the target energy-saving policy. The target energy-saving policy is an energy-saving policy in a policy iteration optimization process. The energy-saving policy includes at least one of the following: changing the activation state of each intra-cluster cell, and adjusting the target parameters of the intra-cluster cells under the condition of meeting the user perception condition. The target parameters are used to adjust the coverage range of the intra-cluster cells;

[0024] A policy evaluation module, configured to evaluate the simulation execution result of the target energy-saving policy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving policy. The comprehensive optimization target is constructed based on user perception metrics and energy-saving metrics;

[0025] A policy execution module, configured to execute the target energy-saving policy when the evaluation result of the target energy-saving policy reaches the optimization completion condition.

[0026] In a possible implementation manner, the policy evaluation module is configured to,

[0027] When the user perception metric determined based on the simulation execution result reaches the user perception condition, calculate a reward value of the target energy-saving policy through a reward function. The reward function is used to evaluate the energy-saving metric;

[0028] When the user perception metric determined based on the simulation execution result does not reach the user perception condition, set the reward value of the target energy-saving policy to a target value.

[0029] In a possible implementation manner, the user perception metric includes at least one of the following: residence ratio metric, reference signal received power metric, transmission rate metric, and interference value metric.

[0030] In a possible implementation manner, the device further includes:

[0031] A policy optimization module, configured to perform policy optimization according to the reward value of the target energy-saving policy when the evaluation result of the target energy-saving policy does not reach the optimization condition until the evaluation result reaches the optimization completion condition.

[0032] In a possible implementation manner, the performance metrics of each intra-cluster cell include the number of cell edge users and the co-coverage of frequency bands. The device further includes:

[0033] A coverage calculation module for separately calculating the weak coverage of each cell within the target cell cluster to obtain the weak coverage of each cell within the cluster;

[0034] A user quantity determination module for determining the cell-edge user quantity of each cell within the cluster based on the weak coverage of each cell within the cluster, where the cell-edge user quantity is used to indicate the potential edge user quantity of the neighboring cells corresponding to the cell within the cluster;

[0035] An analysis module for performing co-coverage analysis of frequency bands for each cell within the cluster to obtain the co-coverage situation of frequency bands for each cell within the cluster.

[0036] In a possible implementation manner, when the target cell within the cluster has edge users, neighboring cells with co-covered frequency bands, and meets the user perception conditions, the target energy-saving policy instructs to adjust the target parameters of the target cell within the cluster and the neighboring cells of the target cell within the cluster; the target cell within the cluster is any cell within the target cell cluster;

[0037] When the target cell within the cluster does not have neighboring cells with co-covered frequency bands, the target energy-saving policy instructs to skip the parameter adjustment operation for the target cell within the cluster.

[0038] On the other hand, a computer device is provided, which includes a processor and a memory. The memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the above-mentioned cell network optimization method.

[0039] On the other hand, a computer-readable storage medium is provided, in which at least one computer program is stored, and the computer program is loaded and executed by a processor to implement the above-mentioned cell network optimization method.

[0040] On the other hand, a computer program product is provided, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer is caused to execute to implement the cell network optimization method provided in the above various optional implementation manners.

[0041] The technical solution provided by this application may include the following beneficial effects:

[0042] The cell network optimization method provided by the embodiments of the present application obtains the performance metric matrix of the target cell cluster, simulates and executes the energy-saving strategy in the policy iteration optimization process based on the performance metric matrix of the target cell cluster. The energy-saving strategy changes the activation states of the cells within each cluster and, under the condition of meeting the user perception condition, adjusts the target parameters for adjusting the coverage range corresponding to the cells within the cluster. The simulation execution result of the target energy-saving strategy is evaluated based on a comprehensive optimization target, which is comprehensively constructed based on user perception metrics, intra-cluster energy consumption, traffic volume, and reference signal received power. When the evaluation result of the target energy-saving strategy reaches the optimization completion condition, the target energy-saving strategy is implemented and executed. Through the above method, by using the performance matrix of the cell cluster and the iterative optimization energy-saving strategy, more accurate decisions can be made in a complex communication environment, effectively adapting to real-time network changes. At the same time, by comprehensively considering multiple factors such as user perception for strategy adjustment and evaluation, the user perception experience is ensured while saving energy, realizing the effective optimization of base station energy saving and cell load, and improving the operation efficiency and service quality of the communication network.

[0043] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present application and used together with the specification to explain the principles of the present application.

[0045] Figure 1 It shows a flowchart of the cell network optimization method provided by an exemplary embodiment of the present application;

[0046] Figure 2 It shows a flowchart of the cell network optimization method provided by another exemplary embodiment of the present application;

[0047] Figure 3 It shows a block diagram of the cell network optimization device provided by an exemplary embodiment of the present application;

[0048] Figure 4 It shows a block diagram of the structure of a computer device shown by an exemplary embodiment of the present application;

[0049] Figure 5 It shows a block diagram of the structure of a computer device shown by another exemplary embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] Exemplary embodiments will be described in detail herein, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0051] An embodiment of the present application provides a method for optimizing a cell network. It can achieve intelligent customization of network energy-saving strategies on the premise of realizing a dynamic balance between energy saving and user perception. Figure 1 The flowchart of the cell network optimization method provided by an exemplary embodiment of the present application is shown. This method can be executed by a computer device, and the computer device can be implemented as a server or a terminal, such as Figure 1 As shown, the cell network optimization method may include the following steps.

[0052] Step 110, obtain a performance index matrix of a target cell cluster, and each vector in the performance index matrix is used to indicate the performance index of each in-cluster cell in the target cell cluster.

[0053] The target cell cluster is any one of multiple cell clusters obtained by clustering based on the base station locations.

[0054] The computer device can obtain the real-time performance index data of each cell in the target cell cluster from a network management system (NMS, Network Management System) or a base station controller (BSC, Base Station Controller).

[0055] Among them, the process of obtaining multiple cell clusters by clustering based on the base station locations can be implemented as:

[0056] Obtain the longitude and latitude information of each base station;

[0057] Perform base station clustering based on the longitude and latitude information of each base station to obtain multiple cell clusters, and each cell cluster can include multiple in-cluster cells with close geographical locations.

[0058] The longitude and latitude information of the base station can include the precision coordinates and dimension coordinates of the base station; in order to improve the accuracy of base station clustering, after obtaining the longitude and latitude information of each base station, the computer device can preprocess the collected data, and the preprocessing includes operations such as data cleaning, feature extraction, and format conversion.

[0059] When performing base station clustering, the computer device can determine the number of cell clusters by comparing the silhouette coefficients. For example, it can determine the optimal number of clusters based on maximizing the silhouette coefficient. Among them, the silhouette coefficient is an index used to evaluate the performance of clustering algorithms and determine the optimal number of clusters. It combines two aspects: cohesion and separation to measure the quality of clustering. For each sample point i in the dataset, the calculation formula for its silhouette coefficient s(i) is as follows:

[0060]

[0061] Among them, a(i) represents the average distance between sample point i and other sample points in the same cluster, that is, the sum of the distances from sample point i to all other points in its cluster divided by the number of sample points in the cluster minus 1. It measures the cohesion within the cluster. The smaller the value of a(i), the closer the distance between sample point i and other points in the cluster, and the higher the cohesion within the cluster. b(i) represents the minimum value of the average distance between the sample point and sample points in other clusters, that is, the minimum value among the average distances from sample point i to each other cluster. It measures the separation between clusters. The larger the value of b(i), the farther the distance between sample point i and other clusters, and the higher the separation between clusters. The silhouette coefficient of the entire dataset is the average of the silhouette coefficients of all sample points, and its value range is between [-1, 1]. A silhouette coefficient close to 1 indicates that the distance between sample point i and the points within its own cluster is much smaller than the distance to other clusters, indicating that the clustering effect is very good, and sample point i is accurately divided into the appropriate cluster. Close to 0 indicates that the distance between sample point i and the points within the cluster is similar to the distance to other clusters, indicating that this sample point may be near the boundary of two clusters, and the clustering result may not be very accurate or there is ambiguity. Close to -1 indicates that sample point i is closer to other clusters and may be misclustered into the current cluster, and the clustering effect is poor.

[0062] When determining the optimal number of clusters based on maximizing the silhouette coefficient, different numbers of clusters k are usually tried, the silhouette coefficient is calculated for each k value, and the k value corresponding to the maximum silhouette coefficient is selected as the optimal number of clusters, so as to obtain a clustering scheme that makes the clustering result more balanced in terms of cohesion and separation.

[0063] After determining the number of cell clusters, the computer device can perform base station clustering based on the longitude and latitude information of each base station through a clustering algorithm to obtain the corresponding number of cell clusters. Among them, the clustering algorithm can be the K-means clustering algorithm, the hierarchical clustering algorithm, etc.

[0064] The performance index matrix of the cell cluster can be a multi-dimensional matrix, where each row or each column represents the performance index of the cells within a cluster. Schematically, assuming there are N cells within a cluster, the performance index of each cell is represented by a vector S iIndicates that it contains M performance metrics, then S i =[s i1 , s i2 , ···, s im , and the performance metric matrix S of this cell cluster can be expressed as:

[0065]

[0066] Among them, the performance metrics of the cells within the cluster can include, but are not limited to, information such as the number of users within the cell coverage, signal strength, interference level, bit error rate, etc. Since the communication network is in real-time change, factors such as user behavior and network load are changing at any time. Therefore, the data of this performance metric matrix can be updated in real-time or regularly.

[0067] Step 120, based on the performance metric matrix of the target cell cluster, simulate and execute the target energy-saving strategy to obtain the simulation execution result of the target energy-saving strategy. The target energy-saving strategy is the energy-saving strategy in the policy iteration optimization process. The energy-saving strategy includes at least one of the following: changing the activation status of each cell within the cluster, and, under the condition of meeting the user perception condition, adjusting the target parameters of the cells within the cluster; the target parameters are used to adjust the coverage range of the cells within the cluster.

[0068] During the policy iteration optimization process, multiple energy-saving strategies can be generated. Each energy-saving strategy contains energy-saving actions, that is, changing the activation status of each cell within the cluster, and, under the condition of meeting the user perception condition, adjusting the target parameters of the cells within the cluster.

[0069] Among them, changing the activation status of the cells within the cluster can include turning off or putting the activation status of the cells within the cluster into sleep, or restoring the activation status of the cells within the cluster; the computer device can determine the changes in network load and communication requirements according to the performance metric matrix of the target cell cluster; in the embodiments of this application, the computer device can introduce a binary state variable If cell n within the cluster is turned off, then Otherwise The set of activation statuses of the cells within the cluster is When the network load is high or the communication requirements in some areas are large, more cells within the cluster are activated to provide more coverage and capacity support. Conversely, when the network load is low or the communication requirements in some areas decline, some cells within the cluster are turned off to save resources and reduce energy consumption.

[0070] Under the condition of meeting the user perception condition, adjusting the target parameters of the cells within the cluster can adjust the coverage range of the cells within the cluster, optimize the handover behavior of the user equipment, making it more inclined to enter the cell with lower energy consumption for load optimization. At the same time, it can ensure that the user perception does not decrease significantly, thereby improving the user experience. Among them, the user perception condition is determined corresponding to the user perception index, and the user perception index refers to the factors perceived by the user during the process of using the network service, such as the clarity of voice calls, the smoothness of data transmission, the stability of network connection, etc. Based on different service requirements, the user perception index and the corresponding user perception condition can have different settings, and this application does not limit this.

[0071] Illustratively, when formulating the energy-saving strategy, if it is determined based on the performance index matrix of the target cell cluster that the traffic volume of a cell within the cluster is low and the user perception index is good, it can be set to the sleep state; if the load of a cell within the cluster is high, its target parameters can be adjusted to guide the user to switch to the neighboring low-load cell, etc.

[0072] Step 130, evaluate the simulation execution result of the target energy-saving strategy based on the comprehensive optimization target to obtain the evaluation result of the target energy-saving strategy; the comprehensive optimization target is constructed based on the user perception index and the energy-saving index.

[0073] The comprehensive optimization target is the basis for evaluating the simulation execution result of the target energy-saving strategy, and is constructed based on multiple key factors. These factors are related to each other and reflect the operating conditions of the network and the user experience from different perspectives. In the embodiments of this application, the comprehensive optimization index can be comprehensively constructed based on the user perception index and the energy-saving index; in a possible implementation manner, the energy-saving index can be a comprehensive index composed of the energy consumption within the cluster, the traffic volume, and the reference signal receiving power (RSRP, Reference Signal Receiving Power). Among them, the energy consumption within the cluster refers to the total energy consumption of all cells within the cell cluster, the traffic volume refers to the total amount of various communication services that the cell cluster can carry, and the reference signal receiving power is a key parameter for measuring the strength of the wireless signal, which directly affects the communication quality between the user equipment and the base station. The comprehensive optimization target indicates that under the premise that the user perception index meets the user perception condition, the comprehensive energy-saving effect of the cell cluster is the best, manifested as the energy consumption within the cluster is as small as possible, the traffic volume is as large as possible, and the reference signal receiving power is as large as possible. It should be noted that based on different energy-saving requirements in actual applications, the energy-saving index can also be a comprehensive index composed of other factors, such as based on the signal interference level, throughput, etc. The embodiments of this application take the energy-saving index as a comprehensive index composed of the energy consumption within the cluster, the traffic volume, and the RSRP as an example for illustration.

[0074] The computer device can simulate the execution of the target energy-saving policy to obtain the simulation execution result of the target energy-saving policy. The simulation execution result includes user perception metrics, intra-cluster energy consumption, traffic volume, RSRP, etc. According to the collected data, the evaluation result of the target energy-saving policy is comprehensively calculated.

[0075] Step 140, when the evaluation result of the target energy-saving policy meets the optimization completion condition, execute the target energy-saving policy.

[0076] If the evaluation result meets the optimization completion condition, it indicates that the current target energy-saving policy is effective. By executing the target energy-saving policy, it is possible to optimize the cell load and improve the energy-saving effect on the premise of minimizing the impact on user perception; otherwise, the target energy-saving policy needs to be adjusted and optimized.

[0077] In summary, the cell network optimization method provided by the embodiments of the present application obtains the performance metric matrix of the target cell cluster, simulates the execution of the energy-saving policy in the policy iteration optimization process based on the performance metric matrix of the target cell cluster. The energy-saving policy changes the activation state of each cell within the cluster, and, under the condition of meeting the user perception condition, adjusts the target parameters for adjusting the coverage range corresponding to the cells within the cluster. The simulation execution result of the target energy-saving policy is evaluated based on the comprehensive optimization target, which is comprehensively constructed based on user perception metrics, intra-cluster energy consumption, traffic volume, and reference signal received power. When the evaluation result of the target energy-saving policy meets the optimization completion condition, implement and execute the target energy-saving policy; through the above method, using the performance matrix of the cell cluster and the iterative optimization energy-saving policy, it is possible to make more accurate decisions in a complex communication environment, effectively adapt to real-time network changes. At the same time, by comprehensively considering multiple factors such as user perception for policy adjustment and evaluation, the user perception experience is ensured while saving energy, the effective optimization of base station energy saving and cell load is achieved, and the operation efficiency and service quality of the communication network are improved.

[0078] Figure 2 The flowchart of the cell network optimization method provided by another exemplary embodiment of the present application is shown. This method can be executed by a computer device, which can be implemented as a server or a terminal, such as Figure 2 shown, the method may include the following steps.

[0079] Step 210, obtain the performance metric matrix of the target cell cluster, where each vector in the performance metric matrix is used to indicate the performance metrics of each cell within the target cell cluster.

[0080] In the embodiments of the present application, the performance metrics of each cell within the cluster may include the number of cell-edge users and the co-coverage of frequency bands. Therefore, when obtaining the performance metrics of each cell within the cluster, the method further includes:

[0081] Calculate the weak coverage degree for each intra-cluster cell in the target cell cluster to obtain the weak coverage degree of each intra-cluster cell.

[0082] Based on the weak coverage degree of each intra-cluster cell, determine the number of cell-edge users for each intra-cluster cell. The number of cell-edge users is used to indicate the potential number of edge users in the neighboring cells corresponding to the intra-cluster cell.

[0083] Conduct frequency band co-coverage analysis for each intra-cluster cell to obtain the frequency band co-coverage situation of each intra-cluster cell.

[0084] Among them, weak coverage refers to the situation where the signal strength in some areas of the network is lower than the normal level, resulting in a decline in communication quality. Edge users refer to the users who will migrate from the serving cell to the neighboring cell after the serving cell performs energy-saving operations. The migration target of edge users can be the cell with the largest RSRP in the neighboring cells.

[0085] When determining the number of cell-edge users, the computer device can extract MR (Measurement Report) data within a certain period. The MR data contains the measurement results of the communication device on the wireless signal at each MR sample point. Based on determining the number of sample points in the coverage area of the intra-cluster cell and the number of sample points in the weak coverage area of the intra-cluster cell from it, the ratio of the two sample point numbers is determined as the weak coverage degree. Among them, the weak coverage area can be the area where the maximum RSRP of the neighboring cell is less than the target power threshold. The value of this target power threshold can be set differently based on the determination criteria for weak coverage. Schematically, this target power threshold can be -110 dBm, etc. Taking the calculation process of the number of cell-edge users of intra-cluster cell i as an example, filter out the sample points with the serving cell being the same intra-cluster cell i from the MR data to form the sample point set corresponding to intra-cluster cell i. Denote the total number of MR sample points in the sample point set as M, and denote the number of MR sample points with the maximum RSRP of the neighboring cell less than the target power threshold in the sample point set as N. The formula for calculating the weak coverage degree can be expressed as:

[0086] Weak coverage degree:

[0087] The formula for calculating the number of cell-edge users of the serving cell can be expressed as:

[0088] E = R × C

[0089] Among them, E represents the number of edge users, R represents the number of RRC (Radio Resource Control) connections of the serving cell during this period, and C represents the weak coverage degree.

[0090] The number of edge users in the primary cell is the potential number of edge users in the neighboring cell.

[0091] When performing frequency band co-coverage analysis on the cells within a cluster, the computer device can determine the frequency band co-coverage situation based on the distance and azimuth angle. Among them, frequency band co-coverage can indicate co-coverage of different radio access technologies or co-coverage of different frequency bands within the same radio access technology. By way of illustration, co-coverage of different radio access technologies can be co-coverage of 5G and 4G, or co-coverage of 4G and 2G, etc. Co-coverage of different frequency bands within the same radio access technology can be co-coverage of different frequency bands under the 4G radio access technology, or co-coverage of different frequency bands under the 5G radio access technology, etc.

[0092] To ensure sufficient coverage overlap between cells and achieve smooth handover and seamless user experience, the computer device can set a distance threshold and an azimuth angle difference threshold. The values of the distance threshold and the azimuth angle difference threshold can be restricted differently based on different co-coverage judgment criteria. By way of illustration, taking the co-coverage of 5G and 4G as an example, with a frequency band distance threshold of 50 meters and an azimuth angle difference threshold of 15 degrees, when D ij <50 and A ij <15, then the 5G cell i and the 4G cell j belong to the co-covered 5G / 4G pair. Among them, D ij represents the distance between the 5G cell i and the 4G cell j, and A ij represents the azimuth angle difference between the 5G cell i and the 4G cell j. When D ij >50 or A ij >15, then the 5G cell i and the 4G cell j do not belong to the co-covered 5G / 4G pair.

[0093] Step 220: Based on the performance metric matrix of the target cell cluster, simulate the execution of the target energy-saving strategy to obtain the simulation execution result of the target energy-saving strategy. The target energy-saving strategy is the energy-saving strategy in the strategy iteration optimization process. The energy-saving strategy includes at least one of the following: changing the activation state of each cell within the cluster, and adjusting the target parameters of the cells within the cluster under the condition of meeting the user perception condition. The target parameters are used to adjust the coverage range of the cells within the cluster.

[0094] In the embodiments of the present application, the goal of iterative optimization of the energy-saving strategy is to achieve energy saving of the cell cluster on the premise of ensuring that the user perception is not significantly affected. To avoid affecting users caused by user migration after the primary cell turns off the activation state, or user migration when balancing the load between cells, during the formulation and iterative optimization of the energy-saving strategy, the frequency band co-coverage situation of the cells within the cluster also needs to be considered. Taking any cell within the cluster in the target cell cluster as an example:

[0095] When there are edge users in the cells within the target cluster, there are neighboring cells with co - coverage of frequency bands, and the user perception conditions are met, the target energy - saving strategy instructs to adjust the target parameters of the cells within the target cluster and the neighboring cells of the cells within the target cluster; the cell within the target cluster is any cell within the target cell cluster.

[0096] When there are no neighboring cells with co - coverage of frequency bands for the cells within the target cluster, the target energy - saving strategy instructs to skip the parameter adjustment operation for the cells within the target cluster.

[0097] That is to say, when formulating the target energy - saving strategy, when the cells within the target cluster simultaneously meet the three conditions of having edge users, having neighboring cells with co - coverage of frequency bands, and meeting the user perception conditions, the target parameters of the cells within the target cluster and the neighboring cells of the cells within the target cluster can be adjusted to switch the edge users of the cells within the target cluster to the neighboring cells. When there are no neighboring cells with co - coverage of frequency bands for the cells within the target cluster, if edge user handover is performed, the user will perceive the handover process, resulting in a poor user experience. Therefore, the parameter adjustment operation for the cells within the target cluster is skipped.

[0098] In a possible implementation, the target parameter can be the CIO (Cell Individual Offset) parameter. By changing the CIO parameter, the signal coverage range of the cell can be changed to optimize network performance. Further, the target parameter can also be the transmit power, antenna tilt angle, handover parameter, etc., which have the function of adjusting the cell coverage range. This application does not limit this.

[0099] Among them, the user perception conditions are formulated based on the corresponding user perception indicators. In a possible implementation, the user perception indicators include at least one of the following: residence ratio indicator, reference signal received power indicator, transmission rate indicator, and interference value indicator. The residence ratio refers to the proportion of the duration that a user device stays in a certain cell in the total statistical duration. For example, within a one - hour statistical time, the user device stays in cell A for 30 minutes, then the residence ratio of the user in cell A is 50%. It should be noted that the statistical duration can be set based on statistical requirements, and this application does not limit this; the reference signal received power is a key parameter used in the LTE network to measure the strength of the wireless signal, which represents the average power of the reference signal received within a certain measurement bandwidth; the transmission rate refers to the amount of data that a user device can transmit per unit time; the interference value refers to the degree of interference caused by other signals except the useful signal to the useful signal during the communication process. The above - mentioned indicators can be used to reflect the user's perception of network handover. Relevant personnel can select one or several of them as user perception indicators based on actual needs and formulate corresponding user perception conditions.

[0100] Schematically, taking the determination of the residence ratio index as a user perception index as an example, in the scenario of 5G / 4G co-coverage, the 5G network duration residence ratio (O t ) is expressed as the following formula:

[0101]

[0102] where O t represents the 5G network duration residence ratio at time t; T 5G_NR represents the residence duration of the 5G terminal in the NR network; T 4G represents the residence duration of the 5G terminal in the 4G network; T 2G represents the residence duration of the 5G terminal in the 2G network.

[0103] This user perception condition can be that the current residence ratio index O t exceeds the residence ratio threshold corresponding to the set residence ratio red line; taking the target parameter as the CIO parameter as an example, on the premise that the current residence ratio index O t exceeds the set residence ratio threshold, the CIO parameter values of the edge user's 5G cell and the corresponding CIO parameter values of the 4G co-coverage cell can be adjusted to improve user perception; in the embodiments of the present application, a parameter variable can be introduced to represent the CIO parameter adjustment value corresponding to the nth cell, with the default value being [0, 0], where the former represents the CIO parameter adjustment value for the primary cell, and the latter represents the CIO parameter adjustment value for the co-coverage cell (i.e., the neighboring cell). Schematically, if the set adjustment value is it means increasing the CIO parameter value of the kth 5G cell by 3 and simultaneously decreasing the CIO parameter value of the corresponding 4G cell covered by this cell by 3, so as to migrate the edge user from the 5G cell to the 4G coverage area to ensure the user experience. It should be noted that the values of the above parameter variables can be adaptively adjusted and set for each cell during the iterative optimization of the energy-saving strategy, and the present application does not limit this.

[0104] The set of adjustment values for each cell in the cell cluster corresponding to the cell cluster can be represented as where each represents the parameter adjustment state of the nth cell.

[0105] After the target energy-saving strategy is simulated and executed, the method further includes:

[0106] Updating the performance index matrix of the target cell cluster based on the simulation execution result of the target energy-saving strategy; that is, updating the performance indexes of each cell in the cluster in the performance index matrix, so that in the subsequent strategy simulation execution, analysis and adjustment can be performed based on the latest network state, thereby continuously optimizing the energy-saving strategy and improving the network performance and energy-saving effect.

[0107] Step 230, evaluate the simulation execution result of the target energy-saving policy based on the comprehensive optimization objective to obtain the evaluation result of the target energy-saving policy; the comprehensive optimization objective is constructed based on user perception indicators, intra-cluster energy consumption, traffic volume, and reference signal received power.

[0108] After changing the activation state and adjusting the target parameters of each intra-cluster cell in the target cell cluster according to the target energy-saving policy, evaluate the simulation execution result according to the network conditions in the updated target cell cluster; in the embodiments of the present application, the computer device can calculate through a reward function to quantitatively obtain the evaluation result; wherein, the reward function can have an initial reward value, and based on the initial reward value, different reward function iterative update calculations are performed based on the satisfaction of user perception conditions, as follows:

[0109] When the user perception indicator determined based on the simulation execution result meets the user perception condition, calculate the reward value of the target energy-saving policy through the reward function; the reward function is used to evaluate the energy-saving indicator;

[0110] When the user perception indicator determined based on the simulation execution result does not meet the user perception condition, set the reward value of the target energy-saving policy to the target value.

[0111] When the user perception indicator determined based on the simulation execution result does not meet the user perception condition, it means that the implementation of the target energy-saving policy will have an adverse impact on the user's communication experience. In this case, in order to reflect the deficiency of the policy, the reward value needs to be adjusted to the target value; wherein, the value of the target value can be a fixed negative feedback value set, or, it can also be determined according to the severity of the non-compliance of the user perception indicator. The value of the target value can be positively correlated with the negative change degree of the target user perception indicator. For example, if the transmission rate indicator in the user perception indicator drops significantly, resulting in a significant deterioration of the user experience, the value of the target value will be relatively large to prompt subsequent policy adjustments to pay more attention to user perception. It should be noted that the transmission rate indicator can be one of the target user perception indicators. Based on actual needs, other indicators can be determined as the target user perception indicators, and the corresponding relationship with the target value can be established. The present application does not limit this.

[0112] Taking the value of the target value as a fixed negative feedback value with a value of 1 and the user perception indicator as the residence ratio indicator as an example, when O t When setting the red line standard, the reward value is:

[0113] r = -1

[0114] When the user perception index reaches the user perception condition, it means that the target energy-saving strategy is timely and has no obvious negative impact on the user's communication experience, or has been improved in some aspects. At this time, the reward value of the target energy-saving strategy is calculated through a reward function, which aims to comprehensively evaluate the energy consumption within the cluster, traffic volume, and RSRP. Schematically, taking the user perception index as the residence ratio index as an example, when O t > the set red line standard, the reward function for calculating the reward value r can be expressed as:

[0115]

[0116] Among them, avg rsrp 、avg U 、avg E are the relevant baseline values recorded before the energy-saving strategy is issued. avg rsrp is the average value of the RSRP of the user's primary cell in all cells within the cluster before the energy-saving strategy is simulated and executed. avg U is the total traffic volume of all cells within the cluster before the energy-saving strategy is simulated and executed. avg E is the total energy consumption value of all cells within the cluster before the energy-saving strategy is issued; mean rsrp is the average value of the RSRP of the user's primary cell in the cells within the cluster that have not performed energy-saving operations after the energy-saving strategy is simulated and executed; U represents the network revenue, which is the total traffic volume of the cells within the cluster that have not performed energy-saving operations after the energy-saving strategy is simulated and executed; E is the energy consumption, which is the total energy consumption value of the cells within the cluster that have not performed energy-saving operations after the energy-saving strategy is simulated and executed; α1, α2, α3 are weight parameters, and the setting range is between 0 and 1. mean rsrp represents the average value of the RSRP of the user's primary cell within the cluster. The larger the value, the smaller the impact of the energy-saving strategy on the user experience; the larger the U value, the smaller the impact of the energy-saving strategy on the traffic volume within the cluster after the energy-saving strategy is simulated and executed, and the energy-saving strategy favors cells with less traffic; the smaller the E value, the smaller the energy consumption within the cluster after the energy-saving strategy is simulated and executed. Among them, α1, α2, α3 are weights, which can be adjusted according to specific requirements: for example, if energy saving is the main optimization goal, the weight of the energy consumption penalty term can be increased to ensure the realization of the energy-saving effect; if traffic load balancing is a more important consideration, the weight of traffic load balancing can be increased to ensure the reasonable allocation of network resources. The reward value r is used as the evaluation standard for the energy-saving strategy and is the weighted value of the above three parameters divided by the baseline value after standardization. The optimization goal of the energy-saving strategy is to make the r value as large as possible.

[0117] When formulating the energy-saving strategy, the priority of user perception is higher than that of energy saving. If the user perception index does not meet the user perception condition in the energy-saving state, r will take a negative feedback target value, such as -1, so as to achieve the purpose of energy saving under the condition that the network duration residence ratio reaches the standard.

[0118] Step 240, when the evaluation result of the target energy-saving strategy reaches the optimization completion condition, execute the target energy-saving strategy.

[0119] Among them, the optimization completion condition can be that after the target strategy is simulated and executed, after experiencing multiple iteration rounds continuously, the reward value corresponding to the evaluation result of the target energy-saving strategy is the highest value, then it is determined that the target energy-saving strategy is the current optimal energy-saving strategy. Among them, the number of multiple iteration rounds can be set according to actual needs, and this application does not limit this; or, this optimization completion condition is that after multiple iteration rounds, the reward value change tends to be stable, and it is determined that the strategy effect tends to be optimal; or, this optimization completion condition can also be that the evaluation result of the target energy-saving strategy indicates that the preset energy-saving index threshold is reached when the user perception index meets the user perception condition. This energy-saving index threshold can be set based on any one or more of the RSRP value, traffic volume, and intra-cluster energy consumption. For example, the average RSRP threshold, the average traffic volume threshold, and the average intra-cluster energy consumption threshold, etc.

[0120] After the evaluation result of the energy-saving strategy reaches the optimization completion condition, the computer device can determine that the energy-saving strategy achieves the balance between network performance and user experience, and issue and execute the energy-saving strategy.

[0121] Step 250, when the evaluation result of the target energy-saving strategy does not reach the optimization condition, optimize the strategy according to the reward value of the target energy-saving strategy until the evaluation result reaches the optimization completion condition.

[0122] In some possible implementation manners, the computer device can perform strategy optimization based on the intelligent algorithm of the reinforcement learning model. Optionally, the reinforcement learning model can include Q learning (Q-Learning) and Boltzmann strategy; select and execute energy-saving actions through the reinforcement learning model to achieve the automatic management and automatic optimization of the base station; among them, the Q value update formula can be expressed as:

[0123] Q(S,A)←Q(S,A)+α[r+γ×max a′ Q(S′,A′)-Q(S,A)]

[0124] α is the learning rate, which determines the weight of new information and old information. γ is the discount factor, which measures the importance of future rewards. r is the immediate reward value obtained by the current action, max A′Q(S′, A′) is the Q - value for selecting the optimal action in the next state S′.

[0125] The Boltzmann strategy sets a relatively high temperature parameter τ at the initial stage. By widely exploring the action space and trying various configuration schemes, it attempts to test the impact of different actions on network performance and user experience. Its exploration formula can be expressed as:

[0126]

[0127] where, P(A i ) is the probability of selecting action A i , Q(S, A i ) is the Q - value of action A i in state S, and τ is the temperature parameter.

[0128] In practical applications, the setting of the temperature parameter τ can gradually decrease over time to ensure more exploration in the initial stage and more utilization of existing experience in the later stage. The adjustment method of the temperature parameter can be expressed as:

[0129]

[0130] where, τ(t) is the temperature parameter at the t - th iteration, τ0 is the initial temperature parameter, β is the annealing rate parameter which determines how fast the temperature drops, and t is the current iteration number. As learning progresses and the temperature parameter τ is gradually decreased, the computer device will be more inclined to select actions with consistently good performance, thereby achieving efficient resource utilization and performance optimization.

[0131] The values of the above - mentioned parameters can be determined through experimental verification using a simulation platform or a virtual environment during algorithm design and testing, so as to reduce interference and risks to the real network and ensure the reliability and security of the algorithm.

[0132] When optimizing the strategy using an intelligent algorithm based on reinforcement learning, the computer device first needs to perform state initialization, including initializing the state space and initializing the reinforcement learning model. Among them, when initializing the state space, the computer device can initialize the state space according to the performance index matrix S of the cell cluster. This performance index matrix includes the number of users within the coverage of each cell in the cluster, signal strength, the number of cell - edge users, and the co - coverage situation of frequency bands, etc. Initializing the reinforcement learning model includes initializing the Q - table, initializing the weight parameters of the reward function, initializing the parameters of the Q - value update formula, and the temperature parameter. When initializing the Q - table, all active states S and energy - saving actions A in Q(S, A) are initialized, and the initial value is set to zero. Initialize the weight parameters α1, α2, α3 of the reward function, the learning rate α, the discount factor γ in the Q - value update formula, and initialize the temperature parameter τ.

[0133] In the process of optimizing for network load balancing and user experience, the computer device selects the optimal energy-saving action A in the current state S according to the Boltzmann strategy. i , and this action selection process includes two parts: activation state adjustment and target parameter adjustment, that is, changing the activation state of some cells within the cluster, and adjusting the target parameters of the cells where the edge users are located after the activation state change and the target parameters of the co-covered cells in the corresponding frequency bands.

[0134] After executing the energy-saving action, the computer device will count the user distribution of the closed cells within the cluster, estimate the change in the traffic volume of the unclosed cells within the cluster where users flow in, calculate the change in the RSRP value of the adjusted user's primary cell, determine the energy consumption and the user perception situation, and then calculate the reward value. Finally, update the relevant metrics based on the reward value to provide a more accurate basis for the next round of policy optimization.

[0135] Taking the energy-saving policy of closing the cell i within the cluster as an example, and after closing the cell i within the cluster, the users in the original cell i within the cluster flow into the cells j and k within the cluster. Assuming there are x users flowing into the cell j within the cluster and y users flowing into the cell k within the cluster, then the traffic volumes U j and U k are changed to:

[0136] U′ j = U j + x

[0137] U′ k = U k + y

[0138] The RSRP values R j and R k of the cells j and k within the cluster are changed to:

[0139] R′ j = f(R j , CIO j )

[0140] R′ k = f(R k , CIO k )

[0141] If there are edge users, then adjust the target parameters to optimize the user experience.

[0142] Taking the CIO parameter as an example of the target parameter, the residence ratio index is changed to:

[0143] O′ t = f(O t , CIO)

[0144] The energy consumption E of cells j and k within a cluster j and E k is changed to:

[0145] E′ j =f(E j , CIO j )

[0146] E′ k =f(E k , CIO k )

[0147] Calculate the reward value based on the RSRP value, traffic volume, energy consumption, and weight parameters after updating according to the energy-saving strategy:

[0148] When O t <the set red line standard:

[0149] r = -1

[0150] When O t ≥ the set red line standard:

[0151]

[0152] Update the Q value according to the reward value:

[0153] Q(S,A) ← Q(S,A) + α[r + γ × max a′ Q(S′,A′) - Q(S,A)]

[0154] Repeat the above process to continuously optimize the energy-saving strategy until the evaluation result of the energy-saving strategy reaches the optimization completion condition, that is, the reinforcement learning model converges, and then issue and execute the energy-saving strategy.

[0155] In a possible implementation, to avoid overfitting problems during the iterative optimization of the energy-saving strategy through the reinforcement learning model, the computer device can use techniques such as experience replay and target network update to reduce the sample correlation during the training process, thereby increasing the generalization ability of the model.

[0156] In addition, considering the convergence and stability of the algorithm in the actual environment, to avoid unstable training processes and results, the computer device can improve the algorithm convergence by adaptively adjusting hyperparameters such as the learning rate and update frequency.

[0157] In another possible application scenario, the comprehensive optimization goal can also be constructed based on multiple metrics such as the energy-saving goal, network throughput, and user perception metrics, so as to balance the relationship between different metrics by constructing a multi-objective reinforcement learning model and achieve more comprehensive network optimization.

[0158] In summary, the cell network optimization method provided by the embodiments of the present application obtains the performance metric matrix of the target cell cluster, and based on the performance metric matrix of the target cell cluster, simulates and executes the energy-saving strategy in the policy iteration optimization process. The energy-saving strategy changes the activation states of the cells within each cluster, and, when the user perception condition is satisfied, adjusts the target parameters corresponding to the cells within the cluster for adjusting the coverage range. The simulation execution result of the target energy-saving strategy is evaluated based on the comprehensive optimization target, which is comprehensively constructed based on user perception metrics, intra-cluster energy consumption, traffic volume, and reference signal received power. When the evaluation result of the target energy-saving strategy reaches the optimization completion condition, the target energy-saving strategy is implemented and executed. Through the above method, by using the performance matrix of the cell cluster and the iterative optimization energy-saving strategy, more accurate decisions can be made in a complex communication environment, effectively adapting to real-time network changes. At the same time, by comprehensively considering multiple factors such as user perception for strategy adjustment and evaluation, the user perception experience is ensured while saving energy, realizing the effective optimization of base station energy saving and cell load, and improving the operation efficiency and service quality of the communication network.

[0159] Figure 3 FIG. shows a block diagram of a cell network optimization device provided by an exemplary embodiment of the present application. The device can be applied in a computer device and execute all or part of the steps of the embodiment as shown in Figure 1 or Figure 2 The computer device can be implemented as a server or a terminal, such as Figure 3 , and the device may include the following modules.

[0160] An acquisition module 310, configured to acquire a performance metric matrix of a target cell cluster, where each vector in the performance metric matrix is used to indicate the performance metrics of each cell within the target cell cluster;

[0161] A policy optimization module 320, configured to simulate and execute a target energy-saving strategy based on the performance metric matrix of the target cell cluster to obtain a simulation execution result of the target energy-saving strategy. The target energy-saving strategy is an energy-saving strategy in the policy iteration optimization process, and the energy-saving strategy includes at least one of the following: changing the activation states of the cells within each cluster, and, when the user perception condition is satisfied, adjusting the target parameters of the cells within the cluster; the target parameters are used to adjust the coverage range of the cells within the cluster;

[0162] A policy evaluation module 330, configured to evaluate the simulation execution result of the target energy-saving strategy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving strategy; the comprehensive optimization target is constructed based on user perception metrics and energy-saving metrics;

[0163] A policy execution module 340, configured to execute the target energy-saving policy when the evaluation result of the target energy-saving policy meets the optimization completion condition.

[0164] In a possible implementation, the policy evaluation module 330 is configured to,

[0165] When the user perception metric determined based on the simulation execution result meets the user perception condition, calculate the reward value of the target energy-saving policy through a reward function; the reward function is used to evaluate the energy-saving metric;

[0166] When the user perception metric determined based on the simulation execution result does not meet the user perception condition, set the reward value of the target energy-saving policy to a target value.

[0167] In a possible implementation, the user perception metric includes at least one of the following: residence ratio metric, reference signal received power metric, transmission rate metric, and interference value metric.

[0168] In a possible implementation, the device further includes:

[0169] A policy optimization module, configured to perform policy optimization according to the reward value of the target energy-saving policy when the evaluation result of the target energy-saving policy does not meet the optimization condition until the evaluation result meets the optimization completion condition.

[0170] In a possible implementation, the performance metrics of each cell in the cluster include the number of cell-edge users and the co-coverage of frequency bands; the device further includes:

[0171] A coverage calculation module, configured to calculate the weak coverage of each cell in the target cell cluster respectively to obtain the weak coverage of each cell in the cluster;

[0172] A user quantity determination module, configured to determine the cell-edge user quantity of each cell in the cluster based on the weak coverage of each cell in the cluster, where the cell-edge user quantity is used to indicate the potential edge user quantity of the neighboring cell corresponding to the cell in the cluster;

[0173] An analysis module, configured to perform frequency band co-coverage analysis on each cell in the cluster to obtain the frequency band co-coverage of each cell in the cluster.

[0174] In a possible implementation, when the target cell in the cluster has edge users, neighboring cells with co-coverage of frequency bands and meets the user perception condition, the target energy-saving policy instructs to adjust the target parameters of the target cell in the cluster and the neighboring cells of the target cell in the cluster; the target cell in the cluster is any cell in the target cell cluster;

[0175] In the case that the cells within the target cluster do not have neighboring cells with co-covered frequency bands, the target energy-saving policy instructs to skip the parameter adjustment operation for the cells within the target cluster.

[0176] In summary, the cell network optimization device provided by the embodiments of the present application obtains the performance index matrix of the target cell cluster, and based on the performance index matrix of the target cell cluster, simulates and executes the energy-saving policy in the policy iteration optimization process. The energy-saving policy changes the activation states of the cells within each cluster, and, under the condition of meeting the user perception condition, adjusts the target parameters for adjusting the cell coverage range in the cells within the cluster. The simulation execution result of the target energy-saving policy is evaluated based on the comprehensive optimization target, which is comprehensively constructed based on the user perception index, the energy consumption within the cluster, the traffic volume, and the reference signal received power. When the evaluation result of the target energy-saving policy reaches the optimization completion condition, the target energy-saving policy is implemented and executed; through the above device, by using the performance matrix of the cell cluster and the iterative optimization energy-saving policy, more accurate decisions can be made in a complex communication environment, effectively adapting to real-time network changes. At the same time, by comprehensively considering multiple factors such as user perception for policy adjustment and evaluation, the user perception experience is ensured while saving energy, the effective optimization of base station energy saving and cell load is realized, and the operation efficiency and service quality of the communication network are improved.

[0177] Figure 4 FIG. shows a block diagram of the structure of a computer device 400 according to an exemplary embodiment of the present application. The computer device may be implemented as the server in the above solution of the present application. The computer device 400 includes a central processing unit (CPU) 401, a system memory 404 including a random access memory (RAM) 402 and a read-only memory (ROM) 403, and a system bus 405 connecting the system memory 404 and the central processing unit 401. The computer device 400 further includes a mass storage device 406 for storing an operating system 409, application programs 410, and other program modules 411. The above system memory 404 and mass storage device 406 may be collectively referred to as a memory.

[0178] According to various embodiments of the present disclosure, the computer device 400 may also run on a remote computer connected to the network through a network such as the Internet. That is, the computer device 400 may be connected to the network 408 through a network interface unit 407 connected to the system bus 405, or in other words, the network interface unit 407 may also be used to connect to other types of networks or remote computer systems (not shown).

[0179] The memory further includes at least one instruction, at least one program, a code set or an instruction set, which are stored in the memory, and the central processing unit 401 implements all or part of the steps in the cell network optimization method shown in the above various embodiments by executing the at least one instruction, at least one program, the code set or the instruction set.

[0180] Figure 5 FIG. shows a structural block diagram of a computer device 500 shown in an exemplary embodiment of the present application. The computer device 500 can be implemented as the above terminal device, such as: a smart phone, a tablet computer, a notebook computer, a desktop computer, etc. The computer device 500 may also be referred to by other names such as a user equipment, a portable terminal, a laptop terminal, a desktop terminal, etc.

[0181] Generally, the computer device 500 includes: a processor 501 and a memory 502.

[0182] In some embodiments, the computer device 500 may further optionally include: a peripheral device interface 503 and at least one peripheral device. The processor 501, the memory 502 and the peripheral device interface 503 can be connected by a bus or a signal line. Each peripheral device can be connected to the peripheral device interface 503 through a bus, a signal line or a circuit board. Specifically, the peripheral devices include at least one of a radio frequency circuit 504, a display screen 505, a camera assembly 506, an audio circuit 507 and a power supply 508.

[0183] In some embodiments, the computer device 500 further includes one or more sensors 509. The one or more sensors 509 include, but are not limited to: an acceleration sensor 510, a gyroscope sensor 511, a pressure sensor 512, an optical sensor 513 and a proximity sensor 514.

[0184] Those skilled in the art can understand that Figure 5 the structure shown in does not constitute a limitation on the computer device 500, and it may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component layout.

[0185] In an exemplary embodiment, a computer-readable storage medium is further provided. At least one computer program is stored in the computer-readable storage medium and is loaded and executed by a processor to implement all or part of the steps in the above-mentioned cell network optimization method. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.

[0186] In an exemplary embodiment, a computer program product is further provided. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions that, when executed by a computer, cause the computer to execute to implement all or part of the steps in the above-mentioned Figure 1 or Figure 2 cell network optimization method shown in the embodiment.

[0187] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only to be regarded as exemplary, and the true scope and spirit of the present application are pointed out by the claims.

[0188] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A method for optimizing a cell network, characterized in that, The method includes: Obtaining a performance metric matrix of a target cell cluster, where each vector in the performance metric matrix is used to indicate the performance metrics of each in-cluster cell in the target cell cluster; Simulating the execution of a target energy-saving policy based on the performance metric matrix of the target cell cluster to obtain a simulation execution result of the target energy-saving policy. The target energy-saving policy is an energy-saving policy in a policy iteration optimization process, and the energy-saving policy includes at least one of the following: changing the activation state of each in-cluster cell, and adjusting the target parameters of the in-cluster cells when the user perception condition is met; the target parameters are used to adjust the coverage range of the in-cluster cells; Evaluating the simulation execution result of the target energy-saving policy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving policy; the comprehensive optimization target is constructed based on user perception metrics and energy-saving metrics; When the evaluation result of the target energy-saving policy meets the optimization completion condition, execute the target energy-saving policy.

2. The method according to claim 1, wherein The evaluating the simulation execution result of the target energy-saving policy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving policy includes: When the user perception metric determined based on the simulation execution result meets the user perception condition, calculating a reward value of the target energy-saving policy through a reward function; the reward function is used to evaluate the energy-saving metric; When the user perception metric determined based on the simulation execution result does not meet the user perception condition, setting the reward value of the target energy-saving policy to a target value.

3. The method according to claim 1 or 2, characterized in that, The user perception metrics include at least one of the following: residence ratio metric, reference signal received power metric, transmission rate metric, and interference value metric.

4. The method according to claim 2, characterized in that, The method includes: When the evaluation result of the target energy-saving policy does not meet the optimization condition, performing policy optimization according to the reward value of the target energy-saving policy until the evaluation result meets the optimization completion condition.

5. The method according to claim 1, wherein The performance metrics of each in-cluster cell include the number of cell-edge users and the frequency band co-coverage situation; the method further includes: Calculating the weak coverage degree of each in-cluster cell in the target cell cluster respectively to obtain the weak coverage degree of each in-cluster cell; Based on the weak coverage degree of each in-cluster cell, determining the number of cell-edge users of each in-cluster cell respectively, where the number of cell-edge users is used to indicate the potential number of edge users of the neighboring cells corresponding to the in-cluster cell; Performing frequency band co-coverage analysis on each in-cluster cell to obtain the frequency band co-coverage situation of each in-cluster cell.

6. The method according to claim 5, wherein When a target in-cluster cell has edge users, has neighboring cells with frequency band co-coverage, and meets the user perception condition, the target energy-saving policy instructs to adjust the target parameters of the target in-cluster cell and the neighboring cells of the target in-cluster cell; the target in-cluster cell is any in-cluster cell in the target cell cluster; When the target in-cluster cell does not have neighboring cells with frequency band co-coverage, the target energy-saving policy instructs to skip the parameter adjustment operation of the target in-cluster cell.

7. A cell network optimization device, characterized in that, The apparatus includes: An acquisition module, configured to acquire a performance metric matrix of a target cell cluster, where each vector in the performance metric matrix is used to indicate the performance metrics of each intra-cluster cell in the target cell cluster; A policy optimization module, configured to simulate and execute a target energy-saving policy based on the performance metric matrix of the target cell cluster to obtain a simulation execution result of the target energy-saving policy. The target energy-saving policy is an energy-saving policy in a policy iteration optimization process. The energy-saving policy includes at least one of the following: changing the activation state of each intra-cluster cell, and adjusting the target parameters of the intra-cluster cell under the condition of meeting the user perception condition. The target parameters are used to adjust the coverage range of the intra-cluster cell; A policy evaluation module, configured to evaluate the simulation execution result of the target energy-saving policy based on a comprehensive optimization target to obtain an evaluation result of the target energy-saving policy. The comprehensive optimization target is constructed based on a user perception index and an energy-saving index; A policy execution module, configured to execute the target energy-saving policy when the evaluation result of the target energy-saving policy meets the optimization completion condition.

8. A computer device, characterized in that, The computer device includes a processor and a memory. The memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the cell network optimization method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, At least one computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by a processor to implement the cell network optimization method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer device, the computer device is caused to execute to implement the cell network optimization method according to any one of claims 1 to 6.

Images