Method, equipment, device and medium for cell performance evaluation

Abstract

The embodiment of the invention relates to a cell performance evaluation method, equipment, a device and a medium. The method comprises: obtaining a plurality of cell state samples and a plurality of cell performance samples of at least one cell of a communication network in a plurality of periods, each cell state sample indicating values of a plurality of state indexes of the at least one cell in a corresponding period, each cell performance sample indicates values of a plurality of performance indexes of at least one cell in a corresponding period; clustering the plurality of cell state samples according to the similarity to obtain at least one cell state category; determining cell performance samples respectively corresponding to each cell state category in the at least one cell state category under the plurality of cell configurations from the plurality of cell performance samples; and determining a sorting result of the plurality of cell configurations under the at least one cell state category based on the cell performance samples respectively corresponding to each cell state category in the at least one cell state category under the plurality of cell configurations through a multi-objective optimization algorithm.

Description

Technical Field

[0001] Example embodiments of the present disclosure relate to the field of communications, and more particularly, to methods, devices, apparatuses, and computer-readable media for cell performance evaluation. Background Art

[0002] Cell-based performance monitoring and evaluation are the primary methods for assessing wireless network performance. Typically, dozens or even hundreds of performance-related metrics or parameters need to be monitored and evaluated. Network performance is often influenced by mobile user traffic patterns, such as the number of users per cell, user distribution, and traffic volume. These traffic patterns are constantly changing and may be affected by seasonal factors. Summary of the Invention

[0003] In a first aspect of the present disclosure, a method for cell performance evaluation is provided. The method includes: obtaining multiple cell status samples and multiple cell performance samples of at least one cell of a communication network within multiple periods, each cell status sample indicating the value of multiple status indicators of at least one cell within the corresponding period, each cell performance sample indicating the value of multiple performance indicators of at least one cell within the corresponding period, and multiple cell performance samples being determined when multiple cell configurations are respectively applied; clustering the multiple cell status samples according to similarity to obtain at least one cell status category, each cell status category including at least two cell status samples from the multiple cell status samples; determining, from the multiple cell performance samples, the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations; and determining, through a multi-objective optimization algorithm, a ranking result of multiple cell configurations under at least one cell status category based on the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations.

[0004] In a second aspect of the present disclosure, a network device is provided. The device includes: at least one processor; and at least one memory coupled to the at least one processor, the at least one memory including instructions stored therein, the at least one memory and the instructions being further configured to, together with the at least one processor, cause the device to perform the method according to the first aspect.

[0005] In a third aspect of the present disclosure, a device for cell performance evaluation is provided. The device includes: a component for obtaining multiple cell status samples and multiple cell performance samples of at least one cell of a communication network within multiple periods, each cell status sample indicates the value of multiple status indicators of at least one cell within a corresponding period, each cell performance sample indicates the value of multiple performance indicators of at least one cell within a corresponding period, and multiple cell performance samples are determined when multiple cell configurations are respectively applied; a component for clustering multiple cell status samples according to similarity to obtain at least one cell status category, each cell status category includes at least two cell status samples from multiple cell status samples; a component for determining, from the multiple cell performance samples, the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations; and a component for determining, through a multi-objective optimization algorithm, the ranking results of multiple cell configurations under at least one cell status category based on the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations.

[0006] In a fourth aspect of the present disclosure, a computer-readable medium is provided, wherein instructions are stored on the computer-readable medium, and when the instructions are executed by at least one processing unit, the at least one processing unit is configured to perform the method according to the first aspect.

[0007] It should be understood that the contents described in the Summary of the Invention section are not intended to limit the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Exemplary embodiments of the present disclosure are presented by way of example, and their advantages are explained in more detail below with reference to the accompanying drawings, in which

[0009] Figure 1 A schematic diagram illustrating an environment in which example embodiments described in this disclosure may be implemented;

[0010] Figure 2 A schematic diagram illustrating an architecture for cell performance evaluation according to some example embodiments of the present disclosure is shown;

[0011] Figure 3 A schematic diagram illustrating a process for cell performance evaluation according to some example embodiments of the present disclosure is shown;

[0012] Figure 4 A schematic diagram illustrating a cell state indicator and a cell performance indicator according to some example embodiments of the present disclosure is shown;

[0013] Figure 5A schematic diagram illustrating average link clustering according to some example embodiments of the present disclosure is shown;

[0014] Figure 6 A schematic diagram illustrating hierarchical clustering according to some example embodiments of the present disclosure is shown;

[0015] Figure 7 A schematic diagram illustrating KPI comparison based on the Pareto Front (PF) algorithm according to some example embodiments of the present disclosure is shown;

[0016] Figure 8 A flowchart illustrating a process for cell performance evaluation according to some example embodiments of the present disclosure is shown;

[0017] Figure 9 shows a simplified block diagram of a device suitable for implementing an example embodiment of the present disclosure; and

[0018] Figure 10 A schematic diagram illustrating a computer-readable medium according to some example embodiments of the present disclosure is shown.

[0019] Throughout the drawings, the same or similar reference numerals denote the same or similar elements. DETAILED DESCRIPTION

[0020] The principles and spirit of the present disclosure will be described below with reference to several exemplary embodiments shown in the accompanying drawings. It should be understood that these specific exemplary embodiments are described only to enable those skilled in the art to better understand and implement the present disclosure, and are not intended to limit the scope of the present disclosure in any way.

[0021] As used herein, the terms "including" and similar terms should be understood as open inclusion, i.e., "including but not limited to." The term "based on" should be understood as "based, at least in part, on." The terms "one embodiment" or "the embodiment" should be understood as "at least one embodiment." The terms "first," "second," etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0022] As used herein, the term "determine" encompasses a wide variety of actions. For example, "determine" may include computing, calculating, processing, deriving, investigating, searching (e.g., searching in a table, database, or another data structure), ascertaining, etc. Furthermore, "determine" may include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), etc. Furthermore, "determine" may include resolving, selecting, choosing, establishing, etc.

[0023] Herein, unless explicitly stated otherwise, executing a step “in response to A” does not mean executing the step immediately after “A” but may include one or more intermediate steps.

[0024] The term "circuitry" as used herein refers to one or more of the following: (a) a hardware circuit implementation only (such as an implementation of analog and / or digital circuitry only); and (b) a combination of hardware circuitry and software, such as (where applicable): (i) a combination of analog and / or digital hardware circuitry and software / firmware, and (ii) any portion of a hardware processor and software (including a digital signal processor, software, and memory that work together to enable a device, such as an optical communication device or other computing device, to perform various functions); and (c) a hardware circuit and / or processor, such as a microprocessor or portion of a microprocessor, that requires software (e.g., firmware) for operation but may operate without software when no software is required for operation.

[0025] The definition of "circuitry" applies to all uses of this term in this application, including in any claims. As another example, the term "circuitry" as used herein also covers an implementation that is solely a hardware circuit or processor (or multiple processors), or a portion of a hardware circuit or processor, or accompanying software or firmware. For example, if applicable to a particular claim element, the term "circuitry" also covers a baseband integrated circuit or a processor integrated circuit or a similar integrated circuit in an OLT or other computing device.

[0026] As used herein, the term "communication network" refers to a network that complies with any suitable communication standard, such as New Radio (NR), Long Term Evolution (LTE), LTE-Advanced (LTE-A), Wideband Code Division Multiple Access (WCDMA), High Speed Packet Access (HSPA), Narrowband Internet of Things (NB-IoT), etc. In addition, the communication between the terminal device and the network equipment in the communication network can be performed according to any suitable generation of communication protocols, including but not limited to first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G), sixth generation (6G) communication protocols and / or any other protocols currently known or to be developed in the future. The example embodiments of the present disclosure can be applied to various communication systems, including but not limited to terrestrial communication systems, non-terrestrial communication systems, or combinations thereof. In view of the rapid development of the communication field, there will certainly be future types of communication technologies and systems that can be used to implement the present disclosure. It should not be regarded as limiting the scope of the present disclosure to only the aforementioned systems.

[0027] As used herein, the term "network device" refers to a node in a communication network via which a terminal device accesses the network and receives services from it. Depending on the terminology and technology applied, a network device may refer to a base station (BS) or an access point (AP), such as a Node B (NodeB or NB), an evolved NodeB (eNodeB or eNB), a NR NB (also called gNB), a remote radio unit (RRU), a radio head (RH), a remote radio head (RRH), a relay, an integrated access and backhaul (IAB) node, a low-power node such as a femto, a pico, and the like. In some example embodiments, a radio access network (RAN) split architecture includes a central unit (CU) and a distribution unit (DU). An IAB node includes a mobile terminal (IAB-MT) part and a DU part, wherein the IAB-MT part is similar to a user equipment (UE) for a parent node, and the DU part is similar to a base station for a next-hop IAB node.

[0028] It should be noted that the titles of any section / subsection provided herein are not limiting. Various embodiments are described throughout this document, and any type of embodiment may be included under any section / subsection. Furthermore, the embodiments described in any section / subsection may be combined in any manner with any other embodiments described in the same section / subsection and / or in different sections / subsections.

[0029] Figure 1 A schematic diagram illustrates an example environment 100 in which example embodiments described herein may be implemented. Example environment 100 may be part of a communications network. Example environment 100 includes at least a terminal device 110 and a network device 120. Network device 120 is, for example, a base station. If terminal device 110 is within a coverage area 122 of network device 120, it may communicate with other devices (e.g., other terminal devices) via a radio link provided by network device 120. Coverage area 122 may include one or more cells.

[0030] It should be understood that Figure 1 The number of devices and their connections shown in the example environment 100 is merely illustrative and not restrictive. Example environment 100 may include any suitable number of devices configured to implement the example embodiments of the present disclosure. Although not shown, it should be understood that one or more other devices may be deployed in communication environment 100.

[0031] Communications in the example environment 100 may be implemented according to any suitable communication protocol(s). Examples of communication protocols include, but are not limited to, first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G), sixth generation (6G), and the like cellular communication protocols, wireless local area network communication protocols such as Institute of Electrical and Electronics Engineers (IEEE) 802.11, and / or any other protocols currently known or to be developed in the future. Furthermore, communications may utilize any suitable wireless communication technology, including, but not limited to, code division multiple access (CDMA), frequency division multiple access (FDMA), time division multiple access (TDMA), frequency division duplex (FDD), time division duplex (TDD), multiple input multiple output (MIMO), orthogonal frequency division multiplexing (OFDM), discrete Fourier transform-based spread spectrum orthogonal frequency division multiplexing (DFT-s-OFDM), and / or any other technology currently known or to be developed in the future.

[0032] Traditional cell performance evaluation methods involve acquiring and offline analyzing key performance indicator (KPI) data from base stations. KPI data for each configuration (feature) is averaged over a certain time period to provide an indication of cell performance. Sometimes, it is necessary to manually select several cells with similar cell conditions (such as the number of user devices and signal-to-noise ratio (SINR)) for more accurate performance comparisons.

[0033] Traditional cell performance evaluation methods have some drawbacks. For example, some evaluation methods can only perform offline performance monitoring and evaluation, and are unable to provide online cell performance analysis and feedback to the algorithm to improve overall cell performance.

[0034] Some evaluation methods also ignore the impact of cell state changes and seasonal factors caused by user traffic patterns. For example, they simply average the KPI data of each configuration over a period of time to obtain cell performance, without considering differences and changes in cell state (such as seasonal changes in cell traffic patterns). This significantly reduces the accuracy of cell performance evaluation. Furthermore, these evaluation methods do not use machine learning techniques such as clustering to analyze cell state.

[0035] In addition, some evaluation methods can only evaluate and compare each KPI individually, such as comparing downlink (DL) throughput, uplink (UL) throughput, call drop rate, or handover success rate, without using a comprehensive approach to evaluate cell performance across multiple KPIs.

[0036] In view of the above-mentioned shortcomings, the traditional cell performance evaluation method needs to be improved to provide more comprehensive and accurate cell performance evaluation results.

[0037] This paper proposes a new cell performance evaluation architecture and method. By studying the relationship between cell status and cell performance, it selects cell status KPIs that describe the cell environment and affect cell performance. It compares the performance of cells with similar status and outputs multiple performance evaluation results for cells with different statuses.

[0038] According to various embodiments of the present disclosure, the method for cell performance evaluation includes obtaining multiple cell status samples and multiple cell performance samples of at least one cell of a communication network within multiple periods. Each cell status sample indicates the value of multiple status indicators of at least one cell within the corresponding period. Each cell performance sample indicates the value of multiple performance indicators of at least one cell within the corresponding period. Multiple cell performance samples are determined when multiple cell configurations are applied separately. The method also includes clustering multiple cell status samples according to similarity to obtain at least one cell status category. Each cell status category includes at least two cell status samples from multiple cell status samples. The method also includes determining, from multiple cell performance samples, the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations. The method also includes determining, through a multi-objective optimization algorithm, the ranking results of multiple cell configurations under at least one cell status category based on the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations. In this way, the accuracy and efficiency of cell performance monitoring and evaluation can be improved. The following will refer to Figures 2 to 7 Detailed description.

[0039] Figure 2 Schematic diagram of an architecture 200 for cell performance evaluation according to some example embodiments of the present disclosure is shown. The architecture 200 includes one or more network devices 120, an evaluation device 210, an operations administration and maintenance (OAM) module 220, and a centralized control system 230.

[0040] One or more network devices 120 may be collectively or individually referred to as network devices 120. The network device 120 includes a functional plane 202, a configuration management module 204, and a performance management module 206. The functional plane 202 further includes a control plane and a user plane. The control plane is responsible for managing and controlling the network device, while the user plane is responsible for forwarding and processing network data. The configuration management module 204 is used to collect, modify, back up, or restore device configurations or parameters. The performance management module 206 is used for performance monitoring, fault detection, performance optimization, etc. Key performance indicators (KPIs) are indicators used to measure device performance, such as throughput, call drop rate, packet loss rate, etc. Counters are used to track and record the number of specific events or operations that occur on a device within different time periods.

[0041] Evaluation device 210 can communicate with one or more network devices 120 and collect data such as cell activity, cell configuration, cell status, and cell performance for use in evaluating cell performance. Evaluation device 210 can communicate with centralized control system 230 via OAM module 220 to obtain configuration information and transmit cell performance evaluation reports. In some embodiments, evaluation device 210 can also interact with software modules in network device 120 (e.g., a data link layer (L2) scheduler) to influence performance improvement algorithms and thereby improve cell performance.

[0042] Evaluation device 210 may include a configuration module 211, a policy controller 212, an evaluation module 213, and an interface module 214. Configuration module 211 is used to define KPI data for cell status and cell performance and manage the cell configuration (features or parameter values, etc.) to be evaluated. Policy controller 212 is used to control cell configuration changes based on the cell configuration and calculation results from evaluation module 213 and interact with other modules (such as performance management module 206, configuration management module 204, algorithms, etc.) to obtain KPI data and cell configuration. Interface module 214 is used to communicate with OAM module 220.

[0043] Evaluation module 213 further includes a KPI data collection module 215, a calculation module 216, and a performance evaluation module 217. KPI data collection module 215 is used to collect KPI data on cell status and cell performance, as well as cell configurations. Calculation module 216 is used to calculate data distribution, distance, clustering, etc. Performance evaluation module 217 is used to provide a cell status-configuration-performance table indicating performance analysis results and to rank the performance of different cell configurations based on a multi-KPI optimization method.

[0044] Evaluation device 210 may be deployed in network device 120 and perform performance evaluation on a cell provided by network device 120. Evaluation device 210 may also be deployed in centralized control system 230 to perform performance evaluation on sites or cells controlled by centralized control system 230. In some example embodiments, evaluation device 210 may be independently deployed in or implemented as a device other than network device 120 and centralized control system 230.

[0045] The OAM module 220 may be deployed in, for example, a core network device to implement operations, management, and maintenance of the network device to ensure network stability, reliability, and manageability.

[0046] The centralized control system 230 is used to centrally manage and control various devices and resources in the network. In some example embodiments, the centralized control system 230 may collect performance evaluation reports of various network devices 120 to generate a comprehensive performance report.

[0047] The above reference Figure 2 The architecture 200 for cell performance evaluation is described. It should be understood that the architecture 200 may include more or fewer devices or modules under the premise of implementing the exemplary embodiments of the present disclosure, and the present disclosure does not limit this. Figures 3 to 7 Describe the cell performance evaluation method.

[0048] Figure 3 FIG2 shows a schematic diagram of a process 300 for cell performance evaluation according to some example embodiments of the present disclosure. The process 300 may be implemented at the network device 120 in the environment 100 or at the evaluation device 210 in the architecture 200. The following takes the implementation at the network device 120 as an example and refers to FIG20. Figure 1 Provide a description.

[0049] Before process 300 begins, the cells provided by network device 120 need to be configured. This includes, for example, defining cell status (also referred to as sample set KPI_set1), cell performance (also referred to as sample set KPI_set2), and cell configuration (such as characteristic flags and parameters to be monitored). Other configurations may also be performed, such as a cell status similarity threshold, as further described below.

[0050] In block 310, the network device 120 collects KPI data. Specifically, the network device 120 obtains cell status samples and performance samples for at least one cell within multiple cycles. The cell status samples (hereinafter also referred to as cell status) indicate the values of multiple cell status indicators (hereinafter also referred to as status indicators) within each cycle. The cell performance samples (hereinafter also referred to as cell performance) indicate the values of multiple cell performance indicators (hereinafter also referred to as performance indicators) within each cycle. In other words, one cycle corresponds to one cell status sample and one cell performance sample. Each cell performance sample is collected and determined over a period of time when multiple cell configurations are respectively applied.

[0051] Figure 4A schematic diagram of cell status indicators and cell performance indicators according to some example embodiments of the present disclosure is shown. Cell status indicators include, for example, capacity, UE status (such as SINR), cell load (such as physical resource block (PRB) utilization, data volume, etc.), UE distribution (such as direction of arrival (DoA), distance to the cell, etc.), UE dispersion, etc. Cell performance indicators include, for example, throughput (including DL throughput, UL throughput), retention (such as call drop rate, packet loss rate, etc.), mobility (such as handover success rate), etc. Note that only examples of cell status indicators and cell performance indicators are given here. Other or different indicators can also be configured in actual applications.

[0052] Cell status indicators and cell performance indicators can be considered two distinct KPI categories, with no overlap. Within a cell, the values of cell status indicators typically change based on UE behavior or cell configuration, while the values of cell performance indicators refer to various indicators used to measure cell performance. Cell status is sometimes also referred to as cell context or cell condition.

[0053] Cell status indicators and performance indicators directly or indirectly influence each other. For example, excessive cell load may lead to increased call drop rates. Another example is the concentration of UEs in a certain direction or within a certain distance range, which may lead to reduced coverage. Figure 4 The relationship 400 between the cell state indicator and the cell performance indicator is shown. Changing one of the cell state indicators may have a direct or indirect impact on a certain cell performance indicator.

[0054] In some example embodiments, at least one cell of interest has the same cell type, thereby providing more appropriate data input for subsequent data analysis and sorting. The at least one cell of the same type may be a cell served by the same network device, or may be a cell provided by different network devices. Typically, within a cell of a type, the cell status determines the cell performance to some extent. For example, more capacity (UE#), higher UE status, higher cell load, or narrower UE distribution (applicable only to MU) will result in higher throughput. More capacity, lower UE status, or higher cell load will result in a higher call drop rate. Compared to averaging all KPIs for different configurations over a time period, selecting cells with similar status and comparing their KPIs will be more accurate, while also avoiding the influence of changes in cell traffic patterns and seasonal effects.

[0055] When collecting cell status samples and cell performance samples, a period can be defined as one hour, one day, one month, or one quarter, etc. This can be determined based on data collection requirements.

[0056] In some example embodiments, each cell status sample includes the values of multiple status indicators collected at multiple unit times of the corresponding period. For example, if the period is represented as Cy=1 day and the unit time is represented as t=15 minutes, then one period includes Cy / t=96 collection points. If the cell status sample includes 3 status indicators, the status indicators and their values can be represented as [s1, s2, s3], then one cell status sample includes 96*[s1, s2, s3] values. In this way, the network device 120 can use the period as a sampling interval to collect KPI data. Similarly, in some example embodiments, each cell performance sample includes the values of multiple performance indicators collected at multiple unit times of the corresponding period. For example, if the period is represented as Cy=1 day and the unit time is represented as t=15 minutes, then one period includes Cy / t=96 collection points. If a cell performance sample includes three performance indicators, the performance indicators and their values can be expressed as [s1, s2, s3], and a cell performance sample includes 96*[s1, s2, s3] values. In this way, the network device 120 can use a period as a sampling interval to collect KPI data.

[0057] At block 320, network device 120 analyzes the collected data. Specifically, for a cell status sample corresponding to each period (which indicates the values of multiple status indicators), network device 120 determines a status distribution corresponding to each status indicator based on the values of the status indicators collected over multiple time units within the period. In some example embodiments, the status distribution corresponding to each status indicator includes a mean and a variance calculated based on the values of the status indicator collected over multiple time units within the period.

[0058] Continuing with the above example, network device 120 can calculate the mean and variance of the three state indicators (e.g., s1, s2, and s3) contained in a cell state sample within a period Cy = 1 day at 96 collection points as the main data features. Based on these data features, network device 120 calculates the state distribution (e.g., normal distribution or Gaussian distribution) of each state indicator. That is, the state distribution of state indicator s1, the state distribution of state indicator s2, and the state distribution of state indicator s3 can be obtained. These data features and distributions will be used for data similarity analysis and trend prediction in subsequent steps of process 300.

[0059] In block 330, network device 120 calculates cell state similarity. Specifically, for each cell using the same cell configuration for at least two periods, network device 120 determines the cell state similarity between multiple state distributions for multiple state indicators. Exemplarily, the cell configurations include cell configuration conf1, cell configuration conf2, and cell configuration conf3. Network device 120 collects cell state samples in periods Cy1, Cy2, and Cy3, respectively. Each cell state sample includes state indicator s1, state indicator s2, and state indicator s3. Network device 120 first calculates state distribution g1 corresponding to state indicator s1, state distribution g2 corresponding to state indicator s2, and state distribution g3 corresponding to state indicator s3. Furthermore, for cell configuration conf1, network device 120 calculates the state similarity between the three state indicators within period Cy1, namely, the similarity between state distribution g1 and state distribution g2, the similarity between state distribution g1 and state distribution g3, and the similarity between state distribution g2 and state distribution g3. Similarly, the network device 120 continues to calculate the state similarities between the three state indicators in the cycle Cy2 and the cycle Cy3.

[0060] By using a normal distribution, patterns can be extracted from the data, allowing discovery of all implemented cell states and categories. In some scenarios, manually evaluating the data is impractical. As the network evolves, machine learning methods are needed to continuously learn and adapt to new state models. In some example embodiments, network device 120 may determine cell state similarity based on a fast Hellinger distance method using a normal distribution.

[0061] The Hellinger distance can be used to measure the similarity between two probability distributions. For example, by comparing the KPI data characteristics and distribution of cell status, the Hellinger distance can quickly calculate the similarity of cell status between different periods. This method can identify similar cell statuses, thereby improving the accuracy of performance evaluation.

[0062] Specifically, the network device 120 determines the Hellinger distance between the multiple state distributions of the multiple state indicators corresponding to one of the at least two cycles and the multiple state distributions of the multiple state indicators corresponding to the other cycle. Furthermore, the network device 120 determines the cell state similarity between the two cycles based on the Hellinger distance. Similarly, the network device 120 calculates the Hellinger distance between each cycle to determine the cell state similarity.

[0063] Exemplarily, the cell configuration includes cell configuration conf1, cell configuration conf2, and cell configuration conf3. The network device 120 collects cell status samples in cycles Cy1 to Cy9, respectively. A fast Hellinger distance calculation method based on normal distribution is then used to measure the similarity of cell states. The Hellinger distance is weighted to obtain a statistical distance D. For example, D[1,2] represents the statistical distance between cell configuration conf1 and cell configuration conf2. Table 1 shows an example matrix of statistical distances for three configurations and nine cycles:

[0064] Table 1 Distance matrix

[0065]

[0066]

[0067] Formula (1) shows the determination of cell state similarity based on the normal distribution determined by the mean and variance calculated in block 320:

[0068]

[0069] Among them, H 2 (P, Q) represents the cell status similarity between the first cycle and the second cycle, σ represents the variance calculated based on the values of the status indicator collected at multiple unit times within a cycle, μ represents the mean calculated based on the values of the status indicator collected at multiple unit times within a cycle, P(σ1, μ1) represents the normal distribution corresponding to the first cycle, and Q(σ2, μ2) represents the normal distribution corresponding to the second cycle.

[0070] Continue to refer Figure 3 In block 340, the network device 120 performs cell state clustering. Specifically, the network device 120 clusters the cell state samples according to similarity to obtain at least one cell state category. Each cell state category includes at least two cell state samples.

[0071] In some example embodiments, network device 120 clusters cell status samples based on cell status similarity using a hierarchical clustering algorithm to obtain at least one cell status category. The hierarchical clustering algorithm is an unsupervised machine learning method. Hierarchical clustering algorithms can be divided into two types: agglomerative and divisive. Agglomerative clustering starts with each data point as a separate cluster, then merges the closest clusters by calculating similarity or distance until all data points are merged into a large cluster, forming a hierarchical clustering tree (pedigree graph). In other words, data points are grouped into different clusters using a bottom-up approach.

[0072] In some example embodiments, average link clustering is used to merge clusters. Average link clustering is a cluster merging method used in hierarchical clustering algorithms. Figure 5 FIG. 4 shows a schematic diagram of average link clustering according to some example embodiments of the present disclosure. Figure 5 As shown, cluster C1 includes data points A to D. Cluster C2 includes data points E to G. The cluster distance between clusters C1 and C2 is obtained by calculating the sum of the distances d(i, j) between each pair of data points between clusters C1 and C2 and then averaging the sums. After clustering, each cluster corresponds to a cell status category.

[0073] Specifically, the distance d(i,j) where i represents a data point in cluster C1 and j represents a data point in cluster C2. For example, D(a,g) represents the distance between data point A and data point G. Formula (2) shows how cluster distance is calculated based on average link clustering:

[0074]

[0075] Among them, N c1 represents the number of data points in cluster C1, N c2 Represents the number of data points in cluster C2.

[0076] The network device 120 can use the distance matrix in Table 1 as input and use a hierarchical clustering algorithm to merge similar cycles into clusters corresponding to cell status categories. Each cluster includes cycles with similar cell status, which may come from different cell configurations. After performing hierarchical clustering, a pedigree diagram can be drawn to show the relationship and similarity between clusters. Figure 6 An example 600 employing hierarchical clustering of cell states is described.

[0077] Figure 6 A schematic diagram of hierarchical clustering according to some example embodiments of the present disclosure is shown. For example, for 36 cycles of data, the cell configuration conf1 is used for cycles from 0 to 17, and the cell configuration conf2 is used for cycles from 18 to 35. The preset distance thresholds for clustering similar cell states include distance threshold 1 and distance threshold 2. For example, distance threshold 1 is 0.25, and distance threshold 2 is 0.45.

[0078] Different distance threshold settings will lead to different hierarchical clustering results. For example, when using a distance threshold of 1, four cell status categories can be obtained. Specifically, based on the statistical distance, the closest periods 31 and 34 are merged into one cluster, and the closest periods 13 and 29 are merged into one cluster. Furthermore, period 20 is merged with the merged periods 13 and 29. Further, periods 8 and 20 are merged with the merged periods 13 and 39. Finally, under the condition that the distance threshold 1 is met, these periods are merged into a large cluster. Thus, periods 31, 34, 8, 20, 13, and 29 are clustered to obtain cell status category c1. Similarly, periods 16, 3, 4, 14, and 15 are clustered to obtain cell status category c2; periods 1, 23, 0, 28, 35, 2, 9, and 17 are clustered to obtain cell status category c3; and periods 24, 25, 11, 12, and 18 are clustered to obtain cell status category c4.

[0079] For another example, when a distance threshold of 2 is used, four cell status categories are obtained. Specifically, periods 31, 34, 8, 20, 13, and 29 are clustered to obtain cell status category c1; periods 16, 3, 4, 14, and 15 are clustered to obtain cell status category c2; periods 6, 30, 7, 22, 1, 23, 0, 28, 35, 2, 9, and 17 are clustered to obtain cell status category c5; and periods 24, 25, 11, 12, 18, 10, 21, 19, and 27 are clustered to obtain cell status category c6.

[0080] The above example uses two distance thresholds to obtain six cell status categories. It should be understood that more distance thresholds can be used to obtain more cell status categories, and this disclosure does not limit this.

[0081] Table 2 shows an example of a cell status category-configuration correspondence table obtained based on hierarchical clustering. Although the table only shows periods, each period also represents the corresponding cell status sample and cell performance sample, which are the main contents of the performance evaluation report output in subsequent steps.

[0082] Table 2 Cell status category-configuration correspondence table

[0083]

[0084] As shown in Table 2, each cell status category contains at least one period, which means that the cell status category corresponds to the cell status samples and cell performance samples collected during these periods. Some cell status categories contain periods from cell configuration conf1 and cell configuration conf2, indicating that the corresponding cell configuration is applied in the corresponding period. For example, cell status category c1 includes period 8 under cell configuration conf1 and periods 13, 20, 31, and 34 under cell configuration conf2. Some cell status categories may only include periods of some cell configurations. For example, cell status category c2 only includes periods of cell configuration conf1. This means that no data corresponding to cell status category c2 is collected under cell configuration conf2.

[0085] In some example embodiments, if the network device 120 determines that the number of cell performance samples corresponding to a first cell configuration among multiple cell configurations under a first cell state category is lower than a first threshold, the first cell configuration is excluded from the sorting for the first cell state category. This is because if the number of samples corresponding to a certain cell configuration is too low, then the sorting result for the cell configuration will be unreasonable, and it can be excluded from the sorting. The sorting of cell state categories will be described in detail in subsequent steps of process 300. As shown in Table 2, the cell state category c1 only includes cell performance samples corresponding to period 8 in the cell configuration conf1. If the number of performance samples is, for example, lower than the threshold number 3, the cell configuration can be deleted in the subsequent configuration sorting for the cell state category.

[0086] Additionally and / or alternatively, if the network device 120 determines that the number of cell performance samples corresponding to the second cell configuration in the plurality of cell configurations under the second cell state category is lower than a second threshold, the sorting for the second cell state category is avoided. This is because if the number of samples corresponding to one or some cell configurations under a certain cell state category is too low, it will be impossible to sort all cell configurations under the cell state category. Therefore, the sorting result of the cell configuration corresponding to the cell state category can be ignored. Continuing to refer to Table 2, for cell state category c2, due to the lack of performance samples under cell configuration conf2 (i.e., lower than the preset number threshold), the configurations under cell state category c2 can be avoided from being sorted.

[0087] Continue to refer Figure 3 The above describes the cell status clustering in block 340. In block 350, the network device 120 determines, from the multiple cell performance samples, the cell performance samples corresponding to the cell status categories under the multiple cell configurations. For example, a cell status-configuration-performance correspondence table can be constructed based on Table 2 and the corresponding data, as shown in Table 3 below:

[0088] Table 3 Cell status-configuration-performance table

[0089]

[0090]

[0091] Table 3 shows the cell status samples corresponding to each cell status category and the cell performance samples under each cell configuration obtained in block 340 , wherein KPI [...] is used to represent the performance indicators of interest in the cell performance samples.

[0092] The above example embodiment analyzes the data characteristics and distribution of cell status KPIs within each cycle and calculates the inter-cycle cell status similarity using the fast Hellinger distance method based on normal distribution. Furthermore, an agglomerative clustering algorithm is employed to automatically establish a core framework for cell status clustering based on cell status similarity, thereby obtaining more accurate cell performance for similar cell states.

[0093] In box 360, the network device 120 performs a multi-KPI comparison. Specifically, the network device 120 determines the ranking results of multiple cell configurations under each cell state category based on the cell performance samples corresponding to each cell state category under multiple cell configurations through a multi-objective optimization algorithm. For example, for cell state category c1, based on the comparison results of the KPI data corresponding to the cell configuration conf1 and the cell configuration conf2, the optimal cell configuration under the cell state category is determined. If the KPI data includes only one KPI, the optimal cell configuration can be easily obtained. If the KPI data includes multiple KPIs, it is necessary to compare them through a multi-objective optimization algorithm. The multi-objective optimization algorithm is used to solve optimization problems with multiple objectives (such as multiple KPIs). These objectives may affect each other, making it more difficult to determine the ranking results.

[0094] In some example embodiments, the multi-objective optimization algorithm includes a Pareto Frontier (PF) algorithm. The PF algorithm is used to rank multiple KPIs in cell performance samples between cell configurations. For example, by calculating the distances between multiple KPIs for each cell configuration and the minimum value (denoted as minPt) of each KPI dimension, the cell configuration corresponding to the multiple KPIs with the smallest distance is found.

[0095] Figure 7 A schematic diagram illustrating an example 700 of performing KPI comparison based on a Pareto front algorithm according to some example embodiments of the present disclosure is shown. Figure 7The figure shows a comparison of two performance metrics, K1 and K2, based on the Pareto Front algorithm. K2 is the call drop rate, which measures the probability of call interruption during communication. Therefore, a larger value on the ordinate indicates a higher probability of call interruption. K1 is the inverse of throughput, which represents the amount of data transmitted per unit time. Therefore, a larger value on the abscissa indicates a smaller amount of data transmitted per unit time, i.e., lower throughput. Given given constraints, it is impossible to simultaneously achieve a lower call drop rate and higher throughput, so a trade-off must be made.

[0096] The Pareto front coordinate diagram can show the trade-off between two performance indicators. Figure 7 As shown in the figure, each point in the coordinate graph represents a performance sample under different cell configurations. The cell configuration corresponding to the point M closest to minPt is the most ideal, indicating both a low call drop rate and a high throughput.

[0097] The Pareto Front algorithm can be used to rank multiple cell configurations corresponding to cell status categories. For example, cell statuses include Cs1 through Cs3. Cell configurations include Conf0 through Conf3. The results of ranking cell configurations based on the Pareto Front algorithm and cell performance samples are shown in Table 4.

[0098] Table 4 Cell configuration ranking

[0099]

[0100] As shown in Table 4, for cell state Cs1, cell configuration Conf1 is optimal and cell configuration Conf3 is the worst. For cell states Cs2 and Cs3, cell configuration Conf0 is optimal and cell configuration Conf3 is the worst.

[0101] In some example embodiments, network device 120 may determine an overall ranking result based on the ranking results of multiple cell configurations under multiple cell status categories. As shown in Table 4, the ranking results of each of the three cell statuses may be weighted to obtain a comprehensive ranking result, that is, cell configuration Conf0 is optimal.

[0102] The above example embodiment uses KPIs to represent cell performance and cell status, using data distribution analysis and statistical distance analysis to measure the similarity between probability distributions. Furthermore, a bottom-up approach can be used to group data points into clusters based on similarity, and multi-objective optimization techniques can be used to rank cell performance across multiple KPIs, thereby better evaluating and comparing cell performance across multiple KPIs. Furthermore, by comparing and optimizing cell performance under different configurations, online cell performance monitoring and multi-configuration comparisons can be performed, providing enhanced performance evaluation results.

[0103] Continue to refer Figure 3 At block 304, the network device 120 outputs a performance evaluation report. Such a performance evaluation report may include, for example, the above-mentioned cell configuration ranking result. The performance evaluation report may have various uses in the network, some example uses of which are discussed below.

[0104] In summary, the present disclosure proposes a cell performance evaluation architecture that can be deployed on a base station or on a server, or as a separate device for online cell performance monitoring and multi-configuration comparison. In this solution, an unsupervised learning method, namely agglomerative hierarchical clustering, is adopted. Agglomerative clustering uses a bottom-up approach to group cell state data (periods) into different clusters based on similarity, and then connects these clusters by merging the two most similar clusters together. Similar cell states will be clustered, and then the cell performance in each cell state cluster can be averaged in each configuration and compared between configurations. The performance evaluation results are put into a cell state-configuration-performance table. Finally, the multi-objective optimization method Pareto front is used to analyze the cell performance of multiple KPIs.

[0105] Based on the cell performance evaluation architecture proposed in this disclosure, cell status prediction and policy control can also be performed.

[0106] Continue to refer Figure 3 In block 370, network device 120 performs a cell state prediction. In block 380, network device 120 may implement policies and controls based on the cell state prediction results. In block 390, network device 120 may modify the cell configuration. For example, based on the cell state prediction results and the records in the cell state-configuration-performance table, the cell configuration may be modified to a configuration that does not or rarely occurs. Specifically, network device 120 determines the predicted values of multiple state indicators for a first cell among at least one cell within a given period. Furthermore, network device 120 configures the first cell using a given cell configuration based on the predicted values of the multiple state indicators. Furthermore, network device 120 collects the values of multiple performance indicators for the first cell within a given period, determines the predicted values of the multiple state indicators as cell state samples, and determines the collected values of the multiple performance indicators as cell performance samples within the given period. The cell state samples and cell performance samples obtained in this manner may be added to the data collected in block 310. In this way, the cell performance samples can be supplemented, thereby improving the accuracy of subsequent data analysis and sorting, and enhancing the efficiency of performance evaluation.

[0107] In some example embodiments, cell state prediction is triggered after each cycle. The Autoregressive Integrated Moving Average (ARIMA) model can model and predict time series data and is a classic method for time series prediction. This method can be used to predict cycle-based cell states, which can be used by network device 120 (e.g., policy controller 212) to adjust cell configurations and generate more similarities between configurations, which can be used for comparison in subsequent steps.

[0108] The ARIMA model consists of three parts: (1) an autoregressive (AR) model, which indicates that a variable regresses to its own lagged or previous value; (2) a differencing (I) module: used to differentiate the original observations to make the time series stationary (i.e., the data value is replaced by the difference between the data value and the previous value); and (3) a moving average (MA) model, which combines the dependence between the observations and the residual errors of the moving average model applied to the lagged observations. In other words, the ARIMA model can predict future values based on past data and trends. It does this by differentiating the time series data to make it stationary, and then building a mathematical model based on the concepts of autoregression and moving average to capture the autocorrelation and lag effects in the time series data. The ARIMA model can be used to predict future trends, seasonal changes, and cyclical fluctuations.

[0109] In some example embodiments, network device 120 may also provide a cell configuration recommendation for the target cell. For example, network device 120 may change the cell configuration to one with optimal performance based on the cell state prediction results and the records in the cell state-configuration-performance table. Specifically, network device 120 determines the expected cell state of the target cell within a target period and determines a target cell state category that matches the expected cell state. Furthermore, network device 120 may select a target cell configuration based on the ranking results of multiple cell configurations under the target cell state category, and then provide the target cell with a recommendation for utilizing the target cell configuration within the target period. In this way, cell performance in the network can be improved.

[0110] In some example embodiments, network device 120 may also control the collection of KPI data through a KPI list in settings.

[0111] In some example embodiments, network device 120 may also trigger a fault / alarm based on the performance evaluation report. Specifically, if the performance evaluation report determines that there are abnormal results in the cell configuration ranking result or the cell status-configuration-performance table, network device 120 may trigger an abnormal alarm for the cell status category or cell configuration.

[0112] In some embodiments, in a communication network, performance evaluation reports corresponding to different types of cells may be maintained for use in status prediction, configuration recommendation, abnormality alarm, etc. for the corresponding cells.

[0113] In summary, the various embodiments of the present disclosure improve the accuracy and efficiency of cell performance monitoring and evaluation by studying the relationship between cell status and performance, and using methods such as hierarchical clustering and Pareto front. Specifically, the various embodiments of the present disclosure can select appropriate cell status KPIs and compare the performance of cells with similar status, thereby providing more accurate performance evaluation results. At the same time, by analyzing the data characteristics and distribution of cell status KPIs and using similarity calculation methods, a cell status clustering framework can be automatically established to further improve the accuracy of performance evaluation. Furthermore, using the Pareto frontier planning method to perform a comprehensive multi-KPI cell performance comparison, rather than comparing the performance of a single KPI alone, can provide more comprehensive performance evaluation results. In addition, by using cell status prediction results and a policy controller, more cell status similarities can be generated to supplement performance samples or optimize cell configurations.

[0114] Example Method

[0115] Figure 8 800 according to some exemplary embodiments of the present disclosure. The method 800 may be implemented in the exemplary environment 100, for example, at the network device 120. The method 800 may also be implemented in Figure 2 The evaluation device 210 in FIG.

[0116] At block 810, network device 120 obtains multiple cell status samples and multiple cell performance samples for at least one cell of the communications network over multiple periods. Each cell status sample indicates values of multiple status indicators of the at least one cell over a corresponding period. Each cell performance sample indicates values of multiple performance indicators of the at least one cell over a corresponding period. The multiple cell performance samples are determined when multiple cell configurations are respectively applied.

[0117] In block 820, the network device 120 clusters the multiple cell status samples according to similarity to obtain at least one cell status category, wherein each cell status category includes at least two cell status samples from the multiple cell status samples.

[0118] In block 830 , the network device 120 determines, from the multiple cell performance samples, cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations.

[0119] In block 840 , the network device 120 determines a ranking result of multiple cell configurations under at least one cell state category based on cell performance samples corresponding to each cell state category under multiple cell configurations in the at least one cell state category by using a multi-objective optimization algorithm.

[0120] In some example embodiments, each cell status sample includes values of multiple status indicators collected at multiple unit times of a corresponding period, and clustering the multiple cell status samples according to similarity includes: for each period in the multiple periods, determining multiple state distributions of each of the multiple status indicators corresponding to the period based on the values of the multiple status indicators collected at multiple unit times within the period; determining the cell state similarity between the multiple state distributions of each of the multiple status indicators corresponding to at least two periods in at least one cell adopting the same cell configuration among multiple cell configurations; and clustering the multiple cell status samples in the multiple periods by a hierarchical clustering algorithm based on the cell state similarity between the multiple state distributions of each of the multiple status indicators determined for the multiple periods to obtain at least one cell state category.

[0121] In some example embodiments, determining the cell state similarity between multiple state distributions of each of multiple state indicators corresponding to at least two cycles includes: determining the Hellinger distance between the multiple state distributions of each of the multiple state indicators corresponding to a first cycle of at least two cycles and the multiple state distributions of each of the multiple state indicators corresponding to a second cycle of at least two cycles; and determining the cell state similarity between the first cycle and the second cycle based on the Hellinger distance.

[0122] In some example embodiments, the state distribution corresponding to each state indicator includes a mean and a variance calculated based on values of the state indicator collected at multiple unit times within the period.

[0123] In some example embodiments, the multi-objective optimization algorithm comprises a Pareto Front (PF) algorithm.

[0124] In some example embodiments, method 800 further includes at least one of the following: if it is determined that the number of cell performance samples corresponding to a first cell configuration among multiple cell configurations under a first cell status category in at least one cell status category is lower than a first threshold, excluding the first cell configuration from the sorting for the first cell status category; or if it is determined that the number of cell performance samples corresponding to at least one cell configuration among multiple cell configurations under a second cell status category in at least one cell status category is lower than a second threshold, avoiding performing sorting for the second cell status category.

[0125] In some example embodiments, obtaining multiple cell status samples and multiple cell performance samples of at least one cell of a communication network within multiple periods includes: determining expected values of multiple status indicators of a first cell in the at least one cell within a given period; configuring the first cell using a given cell configuration from multiple cell configurations within a given period based on the expected values of the multiple status indicators; collecting values of multiple performance indicators of the first cell within a given period; and determining the expected values of the multiple status indicators as cell status samples within the given period, and determining the collected values of the multiple performance indicators as cell performance samples within the given period.

[0126] In some example embodiments, method 800 further includes: determining an expected cell state of the target cell within a target period; determining a target cell state category in at least one cell state category that matches the expected cell state; selecting a target cell configuration from a plurality of cell configurations based on a ranking result of the plurality of cell configurations under the target cell state category; and providing a recommendation to the target cell to utilize the target cell configuration within the target period.

[0127] In some example embodiments, the at least one cell status category includes a plurality of cell status categories, and the method further comprises determining an overall ranking result of the plurality of cell configurations based on ranking results of the plurality of cell configurations under the plurality of cell status categories.

[0128] In some example embodiments, method 800 further includes: determining whether an abnormal alarm for a cell state category or a cell configuration is triggered based on at least one of the ranking results of the multiple cell configurations under each cell state category or the total ranking results of the multiple cell configurations.

[0129] In some example embodiments, at least one cell is of the same cell type.

[0130] Example device

[0131] In some example embodiments, an apparatus for cell performance evaluation may include components for performing the corresponding steps of method 800. These components may be implemented in any suitable manner. For example, the components may be implemented as circuit devices or software modules.

[0132] In some example embodiments, the apparatus may include: a component for obtaining multiple cell status samples and multiple cell performance samples of at least one cell of a communication network within multiple periods, each cell status sample indicating the values of multiple status indicators of at least one cell within a corresponding period, each cell performance sample indicating the values of multiple performance indicators of at least one cell within a corresponding period, and multiple cell performance samples being determined when multiple cell configurations are applied respectively; a component for clustering the multiple cell status samples according to similarity to obtain at least one cell status category, each cell status category including at least two cell status samples from the multiple cell status samples; a component for determining, from the multiple cell performance samples, the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations; and a component for determining, through a multi-objective optimization algorithm, the sorting results of multiple cell configurations under at least one cell status category based on the cell performance samples corresponding to each cell status category in at least one cell status category under multiple cell configurations.

[0133] In some example embodiments, each cell status sample includes values of multiple status indicators collected at multiple unit times of a corresponding period, and the component for clustering the multiple cell status samples according to similarity includes: a component for determining, for each period in the multiple periods, multiple state distributions of each of the multiple status indicators corresponding to the period based on the values of the multiple status indicators collected at multiple unit times within the period; a component for determining, in at least two periods in which at least one cell adopts the same cell configuration among multiple cell configurations, the cell state similarity between the multiple state distributions of each of the multiple status indicators corresponding to at least two periods; and a component for clustering the multiple cell status samples in the multiple periods by a hierarchical clustering algorithm based on the cell state similarity between the multiple state distributions of each of the multiple status indicators determined for the multiple periods, to obtain at least one cell state category.

[0134] In some example embodiments, a component for determining the cell state similarity between multiple state distributions of each of multiple state indicators corresponding to at least two periods includes: a component for determining the Hellinger distance between multiple state distributions of each of multiple state indicators corresponding to a first period of at least two periods and multiple state distributions of each of multiple state indicators corresponding to a second period of at least two periods; and a component for determining the cell state similarity between the first period and the second period based on the Hellinger distance.

[0135] In some example embodiments, the state distribution corresponding to each state indicator includes a mean and a variance calculated based on values of the state indicator collected at multiple unit times within the period.

[0136] In some example embodiments, the multi-objective optimization algorithm comprises a Pareto Front (PF) algorithm.

[0137] In some example embodiments, the apparatus further includes at least one of the following: a component for excluding the first cell configuration from the sorting for the first cell status category if it is determined that the number of cell performance samples corresponding to the first cell configuration among the multiple cell configurations under the first cell status category in at least one cell status category is lower than a first threshold; or a component for avoiding performing sorting for the second cell status category if it is determined that the number of cell performance samples corresponding to at least one cell configuration among the multiple cell configurations under the second cell status category in at least one cell status category is lower than a second threshold.

[0138] In some example embodiments, a component for obtaining multiple cell status samples and multiple cell performance samples of at least one cell of a communication network within multiple periods includes: a component for determining expected values of multiple status indicators of a first cell in at least one cell within a given period; a component for configuring the first cell using a given cell configuration from multiple cell configurations within a given period based on the expected values of the multiple status indicators; a component for collecting values of multiple performance indicators of the first cell within a given period; and a component for determining the expected values of the multiple status indicators as cell status samples within a given period, and determining the collected values of the multiple performance indicators as cell performance samples within a given period.

[0139] In some example embodiments, the apparatus further comprises: means for determining an expected cell state of the target cell within a target period; means for determining a target cell state category in at least one cell state category that matches the expected cell state; means for selecting a target cell configuration from a plurality of cell configurations based on a ranking result of the plurality of cell configurations under the target cell state category; and means for providing a recommendation to the target cell to utilize the target cell configuration within the target period.

[0140] In some example embodiments, the at least one cell status category includes multiple cell status categories, and the apparatus further includes: means for determining an overall ranking result of the multiple cell configurations based on ranking results of the multiple cell configurations under the multiple cell status categories.

[0141] In some example embodiments, the apparatus further comprises: a component for determining whether an abnormal alarm for a cell status category or cell configuration is triggered based on at least one of the ranking results of multiple cell configurations under each cell status category or the total ranking results of multiple cell configurations.

[0142] In some example embodiments, at least one cell is of the same cell type.

[0143] Example devices and media

[0144] Figure 9 FIG1 is a simplified block diagram of a device 900 suitable for implementing an example embodiment of the present disclosure. Device 900 can be used to implement network device 120 in example environment 100 or network device 120 in architecture 200. As shown, device 900 includes one or more processing units 910, one or more memories 920 coupled to processing units 910, and a communication module 940 coupled to processing unit 910.

[0145] The communication module 940 is configured for bidirectional communication. In some exemplary embodiments, the communication module 940 may include at least one antenna to facilitate communication. In some exemplary embodiments, the communication module 940 may include one or more communication interfaces. A communication interface may represent any interface required to communicate with other network elements.

[0146] The processing unit 910 can be of any type suitable for the local technology network and can include, but is not limited to, one or more of a general-purpose computer, a special-purpose computer, a microcontroller, a digital signal controller (DSP), and a controller-based multi-core controller architecture. The device 900 can have multiple processors, such as application-specific integrated circuit chips, which are time-slave to a clock synchronized with a main processor.

[0147] The memory 920 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, read-only memory (ROM) 1024, erasable programmable read-only memory (EPROM), flash memory, hard disks, compact disks (CDs), digital video disks (DVDs), and other magnetic and / or optical storage. Examples of volatile memories include, but are not limited to, random access memory (RAM) 922 and other volatile memories that do not persist across a power outage.

[0148] Computer program 930 includes computer executable instructions for execution by associated processing unit 910. Computer program 930 may be stored in ROM 924. Processing unit 910 may perform any suitable actions and processes by loading computer program 930 into RAM 922.

[0149] The exemplary embodiments of the present disclosure may be implemented with the aid of a computer program 930, so that the device 900 can execute the instructions described in the reference Figures 2 to 8 Any process of the present disclosure discussed. The example embodiments of the present disclosure may also be implemented by hardware or by a combination of software and hardware.

[0150] In some example embodiments, the computer program 930 may be tangibly embodied in a computer-readable medium that may be included in the device 900 (such as in the memory 920) or other storage device accessible by the device 900. The computer program 930 may be loaded from the computer-readable medium into the RAM 922 for execution. The computer-readable medium may include any type of tangible, non-volatile memory, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc.

[0151] Figure 10 An example of a computer readable medium 1000 in the form of a CD or DVD according to some example embodiments of the present disclosure is shown. The computer readable medium 1000 has a computer program 930 stored thereon.

[0152] In general, various embodiments of the present disclosure may be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software, which may be executed by a controller, microprocessor, or other computing device. Although various aspects of the example embodiments of the present disclosure are shown and described as block diagrams, flow charts, or using some other pictorial representation, it should be understood that the blocks, devices, systems, techniques, or methods described herein may be implemented as, by way of non-limiting example, hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or a controller or other computing device, or some combination thereof.

[0153] The present disclosure also provides at least one computer program product tangibly stored on a computer-readable storage medium. In some example embodiments, the computer-readable storage medium may be non-transitory. The computer program product includes computer-executable instructions, such as instructions included in program modules, which are executed in a device on a real or virtual processor of a target to perform the above-referenced Figure 8 Method 800 is described. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or implement specific abstract data types. In various embodiments, the functionality of the program modules can be combined or divided between program modules as needed. The machine-executable instructions for the program modules can be executed on local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0154] The computer program code for implementing the disclosed method can be written in one or more programming languages. These computer program codes can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device so that the program code, when executed by the computer or other programmable data processing device, causes the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on a computer, partially on a computer, as an independent software package, partially on a computer and partially on a remote computer or entirely on a remote computer or server.

[0155] In the context of the present disclosure, computer program code or related data may be carried by any suitable carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, and the like. Examples of signals may include electrical, optical, radio, acoustic, or other forms of propagated signals, such as carrier waves, infrared signals, and the like.

[0156] A computer-readable medium may be any tangible medium that contains or stores a program for or in connection with an instruction execution system, apparatus, or device. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. More detailed examples of computer-readable storage media include an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0157] In addition, although the operations of the method of the present disclosure are described in a particular order in the accompanying drawings, this does not require or imply that these operations must be performed in this particular order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can change the order of execution. Additionally or alternatively, certain steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps. It should also be noted that the features and functions of two or more devices according to the present disclosure can be embodied in one device. Conversely, the features and functions of a device described above can be further divided into being embodied by multiple devices.

[0158] Although the present disclosure has been described with reference to several specific embodiments, it should be understood that the present disclosure is not limited to the specific embodiments disclosed. The present disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for evaluating cell performance, comprising: Acquiring a plurality of cell state samples and a plurality of cell performance samples of at least one cell of a communication network within a plurality of periods, each cell state sample indicating values of a plurality of state indicators of the at least one cell within a corresponding period, each cell performance sample indicating values of a plurality of performance indicators of the at least one cell within a corresponding period, the plurality of cell performance samples being determined when a plurality of cell configurations are respectively applied; Clustering the multiple cell status samples according to similarity to obtain at least one cell status category, each cell status category including at least two cell status samples from the multiple cell status samples; Determine, from the multiple cell performance samples, cell performance samples corresponding to each cell state category in the at least one cell state category under the multiple cell configurations; as well as By means of a multi-objective optimization algorithm, a ranking result of the multiple cell configurations under the at least one cell state category is determined based on the cell performance samples respectively corresponding to each cell state category in the at least one cell state category under the multiple cell configurations.

2. The method according to claim 1, wherein each cell status sample comprises values of the multiple status indicators collected at multiple unit times in a corresponding period, and wherein clustering the multiple cell status samples according to similarity comprises: For each of the multiple cycles, determining multiple state distributions of the multiple state indicators corresponding to the cycle based on the values of the multiple state indicators collected at multiple unit times within the cycle; determining, in at least two periods in which the at least one cell adopts the same cell configuration among the multiple cell configurations, cell state similarities between the multiple state distributions of the multiple state indicators corresponding to the at least two periods; as well as Based on the cell state similarities between the multiple state distributions of the multiple state indicators respectively determined for the multiple periods, the multiple cell state samples in the multiple periods are clustered by a hierarchical clustering algorithm to obtain the at least one cell state category.

3. The method according to claim 2, wherein determining the cell state similarity between the multiple state distributions of the multiple state indicators corresponding to the at least two periods comprises: determining a Hellinger distance between a plurality of state distributions of each of the plurality of state indicators corresponding to a first period of the at least two periods and a plurality of state distributions of each of the plurality of state indicators corresponding to a second period of the at least two periods; as well as The cell state similarity between the first period and the second period is determined based on the Hellinger distance. 4 . The method according to claim 2 , wherein the state distribution corresponding to each state indicator comprises a mean and a variance calculated based on values of the state indicator collected at multiple unit times within the period. The method of claim 1 , wherein the multi-objective optimization algorithm comprises a Pareto Front (PF) algorithm.

6. The method according to claim 1, further comprising at least one of the following: If it is determined that the number of cell performance samples corresponding to a first cell configuration among the multiple cell configurations under a first cell status category among the at least one cell status category is lower than a first threshold, excluding the first cell configuration from the ranking for the first cell status category; or If it is determined that the number of cell performance samples corresponding to at least one cell configuration among the multiple cell configurations under a second cell status category in the at least one cell status category is lower than a second threshold, avoid performing the sorting for the second cell status category.

7. The method according to claim 1, wherein obtaining a plurality of cell status samples and a plurality of cell performance samples of at least one cell of the communication network within a plurality of periods comprises: Determining expected values of the plurality of status indicators for a first cell among the at least one cell within a given period; configuring the first cell using a given cell configuration from among the multiple cell configurations within the given period based on the estimated values of the multiple state indicators; collecting values of the multiple performance indicators of the first cell within the given period; as well as The expected values of the multiple state indicators are determined as cell state samples within the given period, and the collected values of the multiple performance indicators are determined as cell performance samples within the given period.

8. The method according to claim 1, further comprising: determining an expected cell state of the target cell within a target period; determining a target cell state category that matches the expected cell state among the at least one cell state category; selecting a target cell configuration from the multiple cell configurations based on a ranking result of the multiple cell configurations under the target cell state category; as well as A recommendation is provided to the target cell to utilize the target cell configuration during the target period.

9. The method according to claim 1, wherein the at least one cell status category comprises a plurality of cell status categories, and the method further comprises: Based on the ranking results of the multiple cell configurations under the multiple cell status categories, an overall ranking result of the multiple cell configurations is determined.

10. The method according to claim 1 or 9, further comprising: Based on at least one of the ranking results of the multiple cell configurations under each cell status category or the total ranking results of the multiple cell configurations, it is determined whether an abnormal alarm for the cell status category or the cell configuration is triggered. The method according to claim 1 , wherein the at least one cell is of the same cell type.

12. A cell performance evaluation device, comprising: means for obtaining a plurality of cell state samples and a plurality of cell performance samples of at least one cell of a communication network within a plurality of periods, each cell state sample indicating values of a plurality of state indicators of the at least one cell within a corresponding period, each cell performance sample indicating values of a plurality of performance indicators of the at least one cell within a corresponding period, the plurality of cell performance samples being determined when a plurality of cell configurations are respectively applied; a component for clustering the multiple cell status samples according to similarity to obtain at least one cell status category, each cell status category including at least two cell status samples from the multiple cell status samples; a component for determining, from the multiple cell performance samples, cell performance samples corresponding to each cell status category in the at least one cell status category under the multiple cell configurations; as well as A component for determining a ranking result of the multiple cell configurations under the at least one cell state category based on the cell performance samples corresponding to each cell state category in the at least one cell state category under the multiple cell configurations through a multi-objective optimization algorithm.

13. A network device comprising: at least one processor; as well as At least one memory coupled to the at least one processor, the at least one memory including instructions stored therein, the at least one memory and the instructions being further configured to, together with the at least one processor, cause the device to perform the method according to any one of claims 1 to 11. 14 . A computer-readable medium having instructions stored thereon, which, when executed by at least one processing unit, causes the at least one processing unit to be configured to perform the method according to claim 1 .

Images