SYSTEM, METHOD, AND NON-TRANSITORY COMPUTER-READABLE MEDIUM FOR ANTICIPATING CAPACITY VIOLATIONS IN MOBILE NETWORKS - Patent application

Abstract

A capacity violation is predicted in the mobile network. A key performance indicator (KPI) database is accessed to obtain key performance indicator (KPI) data related to capacity of cells in the mobile network. Based on the KPI data, critical cells and non-critical cells are identified. Critical cells exhibit high utilization that impacts performance, and non-critical cells do not exhibit high utilization. For non-critical cells, a predictive model is applied to identify at least one predefined predicted time window associated with a capacity problem associated with at least one of the non-critical cells. Based on applying the predictive model, a report is generated that identifies actions to be taken to address the capacity problem. Actions from the report are executed to configure the mobile network to address the capacity problem of the critical cell and / or the capacity problem of the non-critical cells that have a predicted capacity problem within one of the predefined predicted time windows.

Description

[Technical field]

[0001] The present specification relates to a system, method, and non-transitory computer-readable medium for predicting capacity violations in mobile networks. [Background technology]

[0002] Mobile networks involve the deployment of cell sites over extended geographic areas. With the number of users and the proliferation of bandwidth-intensive services such as video and music streaming, smart devices, and multiplayer video games, the capacity of a particular cell site increases, often to the point where user experience is adversely affected. Highly utilized cells result in network performance issues such as slower upload / download speeds, latency, and coverage issues.

[0003] Currently, due to the lack of proper network monitoring tools, network operators have no advance notice or knowledge that a capacity violation in the network will occur. For example, a cell at one time is experiencing 50% utilization, but soon the utilization increases to 80% or 90%. Identifying these types of problems and addressing them is difficult and time consuming.

[0004] To improve performance and split the load, new sites, antennas, or technologies are installed blindly or after the fact, however this is inefficient and leads to unproductive use of personnel, software, and hardware resources. Summary of the Invention

[0005] In at least one embodiment, a method for predicting capacity violations in a mobile network includes accessing a KPI database to obtain Key Performance Indicators (KPI) data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, where the critical cells exhibit high utilization rates that affect performance by the critical cells and the non-critical cells do not exhibit high utilization rates; applying a predictive model to identify, for the non-critical cells, at least one predefined prediction time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the predictive model that identifies at least one action to be taken to configure the mobile network to address at least one of the capacity problems of the critical cells or the capacity problems of the non-critical cells having a predicted capacity problem within one of the predefined prediction time windows; and performing the at least one action to configure the mobile network to address at least one of the capacity problems of the critical cells or the capacity problems of the non-critical cells having a predicted capacity problem within one of the predefined prediction time windows.

[0006] In at least one embodiment, a device for predicting capacity violations in a mobile network includes a memory storing computer-readable instructions; and a processor configured to execute the computer-readable instructions to: access a KPI database to obtain key performance indicator (KPI) data related to capacity of cells in the mobile network; identify critical cells and non-critical cells based on the KPI data, where the critical cells exhibit high utilization rates that affect performance by the critical cells and the non-critical cells do not exhibit high utilization rates; apply a predictive model for the non-critical cells to identify at least one predefined prediction time window associated with a capacity problem associated with at least one of the non-critical cells; and generate a report based on applying the predictive model that identifies actions to be taken to configure the mobile network to address at least one of the capacity problems of the critical cells or the capacity problems of the non-critical cells that have predicted capacity problems within one of the predefined prediction time windows.

[0007] In at least one embodiment, a non-transitory computer readable medium having computer readable instructions stored thereon, when executed by a processor, causes the processor to perform operations including accessing a KPI database to obtain key performance indicator (KPI) data related to capacity of cells in a mobile network; identifying critical cells and non-critical cells based on the KPI data, where the critical cells exhibit high utilization rates that affect performance by the critical cells and the non-critical cells do not exhibit high utilization rates; for the non-critical cells, applying a predictive model to identify at least one predefined forecast time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the predictive model that identifies at least one action to be taken to configure the mobile network to address at least one of the capacity problems of the critical cell or the capacity problems of the non-critical cells having a predicted capacity problem within one of the predefined forecast time windows; and performing the at least one action to configure the mobile network to address at least one of the capacity problems of the critical cell or the capacity problems of the non-critical cells having a predicted capacity problem within one of the predefined forecast time windows.

[0008] Aspects of the present disclosure are best understood from the following detailed description when read in conjunction with the accompanying drawings, in which: It should be noted that, in accordance with standard industry practice, various features have not been drawn to scale, and in fact dimensions of various features may be increased or decreased for clarity of illustration. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 illustrates a mobile network according to at least one embodiment.

[0010] [Diagram 2]FIG. 1 illustrates a Key Performance Indicator (KPI) list according to at least one embodiment.

[0011] [Diagram 3] FIG. 1 illustrates an orientation and angle calculation according to at least one embodiment.

[0012] [Figure 4] 4 is a flowchart 400 of a method for identifying critical and non-critical cells according to at least one embodiment. [Figure 4C] Continuing with method flowchart 400.

[0013] [Diagram 5] 1 is a flowchart of a method for handling critical cells without prediction according to at least one embodiment.

[0014] [Figure 6] FIG. 2 illustrates a look-up table for determining how to make a forecasting decision, according to at least one embodiment.

[0015] [Figure 7] 1 is a flowchart of a method for handling non-critical cells using prediction according to at least one embodiment.

[0016] [Figure 8] 1 is a flowchart of a method for taking actions to address identified capacity issues, according to at least one embodiment.

[0017] [Figure 9] FIG. 2 illustrates a predetermined prediction window in accordance with at least one embodiment.

[0018] [Figure 10] FIG. 1 illustrates a capacity planning report according to at least one embodiment.

[0019] [Figure 11] 1 is a flowchart of a method for predicting capacity violations in a mobile network according to at least one embodiment.

[0020] [Figure 12] FIG. 1 is a high-level functional block diagram of a processor-based system according to at least one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] The embodiments described herein illustrate examples of implementing different features of the subject matter provided. To simplify the disclosure, examples of components, values, operations, materials, arrangements, and the like are described below. Of course, these are examples and are not intended to be limiting. Other components, values, operations, materials, arrangements, and the like are contemplated. For example, the formation of a first feature above or on a second feature in the following description includes embodiments in which the first feature and the second feature are formed in direct contact, and also includes embodiments in which an additional feature is formed between the first feature and the second feature such that the first feature and the second feature cannot be in direct contact. In addition, the disclosure repeats reference numbers and / or letters in various examples. This repetition is for the purposes of brevity and clarity, and does not dictate a relationship between the various embodiments and / or configurations described.

[0022] Additionally, spatially relative terms such as "beneath," "below," "lower," "above," "upper," and the like are used herein for ease of description to describe the relationship of one element or feature to another element(s) or feature(s) as shown in the figures. The spatially relative terms are intended to encompass different orientations of the device during use or operation in addition to the orientation shown in the figures. The device may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly.

[0023] The terms "system" and "network" in the embodiments of the present application are used interchangeably. "At least one" means one or more, and "multiple" means two or more. The term "and / or" describes an association relationship between related objects and indicates that three relationships exist. For example, A and / or B indicates that only A is present, both A and B are present, and only B is present, where A and B are singular or plural. The character " / " generally indicates an "or" relationship between related objects. "At least one of" or similar expressions means any combination of items, including any combination of a single item or multiple items. For example, "at least one of A, B, and C" includes A, B, C, AB, AC, BC, or ABC, and "at least one of A, B, or C" includes A, B, C, A and B, A and C, B and C, or A and B and C.

[0024] According to at least one embodiment, a method for predicting capacity violations in a mobile network includes accessing a KPI database to obtain key performance indicator (KPI) data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, where the critical cells exhibit high utilization rates that affect performance by the critical cells and the non-critical cells do not exhibit high utilization rates; applying a predictive model to identify, for the non-critical cells, at least one predefined prediction time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the predictive model, that identifies at least one action to be taken to configure the mobile network to address at least one of the capacity problems of the critical cells or the capacity problems of the non-critical cells having a predicted capacity problem within one of the predefined prediction time windows; and performing the at least one action to configure the mobile network to address at least one of the capacity problems of the critical cells or the capacity problems of the non-critical cells having a predicted capacity problem within one of the predefined prediction time windows.

[0025] Advantages associated with the method for predicting capacity violations in a mobile network include identifying cells that will become important in the near future, reducing capacity and network failures due to current capacity loads, improving network performance, proactively managing highly utilized cells, and identifying potential risks in the network. Predicting future network capacity violations helps in planning network expansion to avoid network failures due to capacity violations.

[0026] FIG. 1 illustrates a mobile network 100 according to at least one embodiment.

[0027] In Figure 1, a mobile communication network couples User Equipment (EU) 110 to a Core Network (CN) 150 via a Radio Access Network (RAN) 120. The RAN 120 connects individual devices such as User Equipment (EU) 110 to other parts of the network, such as the CN 150, via wireless connections. The RAN 120 is responsible for managing radio resources, including strategies and algorithms for controlling power, channel allocation, and data rates.

[0028] The RAN 120 has evolved over time from 3G to 5G, and is implemented in various configurations such as Global System for Mobile Communications (GSM) RAN (GRAN), Evolved GSM Enhanced Data Rates (EDGE) RAN (GERAN), Universal Mobile Telecommunications Service (UMTS) Terrestrial RAN (UTRAN), Evolved UMTS Terrestrial RAN (E-UTRAN), Centralized / Cloud RAN (CRAN), Virtualized RAN (VRAN), and Open RAN (ORAN).

[0029] In the 3G network 121, the RAN 120 includes base stations for cell sites 122, 123, referred to as Node Bs (NBs) 124, 125, and a Radio Network Controller (RNC) 126. The RNC 126 controls and manages the radio transceivers in the Node Bs 124, 125, and also manages operational functions such as handoffs and radio channels. The RNC 126 handles communications with the 3G core network 152.

[0030] In the 4G network 132, the cell sites 130, 131 are implemented using evolved Node Bs for radio base stations (eNodeBs or eNBs) 134, 135. The eNodeBs 134, 135 perform radio access functions equivalent to the combined operations that the Node Bs 124, 125 and the RNC perform in 3G, and can connect to an evolved packet core 154.

[0031] In the 5G network 140, the cell sites 141, 142 are implemented using one of two types of RANs: Next Generation Node B (gNodeB) 144 and Next Generation Evolved Node B (ng-eNB) 146. The ng-eNB 146 is an enhanced version of the 4G eNodeB and connects the 5G UE 110 to a 5G Core Network (5GC) 156 using a 4G LTE air interface. The gNB 144 allows the 5G UE 110 to connect with the 5GC 156 using a 5G NR air interface. The gNB 144 and the ng-eNB 146 are interconnected by an Xn interface. The gNB 144 and the ng-eNB 146 are also connected to the 5GC 156 by an NG interface.

[0032] In 5G, for example, an open RAN environment can be implemented, where the RAN 120 provided by the gNB 144 and the ng-eNB 146 is separated into a Radio Unit (RU) 147, a Distributed Unit (DU) 148, and a Centralized Unit (CU) 149. The RU 147 is where the radio frequency signals are transmitted, received, amplified, and digitized. The RU 147 is located near the antenna or is integrated into the antenna. The DU 148 and the CU 149 are the computing parts of the base station and send the digitized radio signals to the network. The DU 148 is physically located at or near the RU 147, while the CU 149 is often located closer to the core network 150. Different interfaces associated with the open RAN 120 include a fronthaul (FH) between the RU 147 and the DU 148, a midhaul (MH) between the DU 148 and the CU 149, and a backhaul (BH) between the CU 149 and the core network 150.

[0033] The Core Network (CN) 150 connects the RAN 120 to networks 160 such as Public Land Mobile Networks (PLMN), Public Switched Telephone Networks (PSTN), and Packet Data Networks (PDN). The CN 150 provides high-level traffic aggregation, routing, call control / switching, user authentication and charging. The 3G CN 152 includes two different domains: circuit-switched and packet-switched elements. The 4G Evolved Packet Core (EPC) 154 includes four main network elements: Serving Gateway (S-GW), Packet Data Network (PDN) Gateway (P-GW), Mobility Management Entity (MME), and Home Subscriber Server (HSS). The S-GW routes and forwards data packets from the UE and acts as a mobility anchor during inter-eNodeB handover. The P-GW acts as an ingress and egress point from a PDN (Packet Data Network), such as the Internet, to the EPC. The MME manages the UE access network and mobility, as well as the establishment of bearer paths for the User Equipment (UE). The MME is also involved in the bearer activation / deactivation process. The HSS is the master database for a given subscriber, acting as a central repository of information for network nodes. Subscriber related information maintained by the HSS includes user identity, security, location, and subscriber profile. The EPC is connected to external networks including the IP Multimedia Core Network Subsystem (IMS). The 5GC 156 supports new network functions (NFs) associated with the packet core and user data management domains. The 5GC 156 provides a disaggregated network architecture that introduces the Service Based Interface (SBI) and Control Plane and User Plane Separation (CUPS). The 5GC disaggregates the 4G MME into an Access and Mobility Management Function (AMF) and a Session Management Function (SMF). The AMF receives connection and session related information from the UE, but is responsible for handling connection and mobility management tasks. Messages related to session management are forwarded to the SMF.

[0034] The network is managed by a network management system (NMS) 170, which provides several network management functions. According to at least one embodiment, the NMS provides an estimate of capacity violations in the mobile network 100. There is a maximum number of users that can be accommodated by the mobile network 100 before either the quality or performance of the mobile network 100 is adversely affected. The carrying capacity of the mobile network 100 is the total amount of data or voice traffic that the cell sites of the mobile network 100, e.g., cell sites 122, 123, 130, 131, 141, 142, are able to transfer to and from their customers. Wireless data is carried by modulating radio waves. The amount of waves (or amount of spectrum) that a wireless system can modulate per second is called its bandwidth, measured in Hertz (Hz). All else being equal, a signal with a higher bandwidth (i.e., more Hz) can carry more data per second than a signal with a lower bandwidth (i.e., lower Hz). The total amount of data a cell site transfers over a given period of time is related to the speed at which the cell sites 122, 123, 130, 131, 141, 142 transfer data bytes. All things being equal, faster cell sites transfer more bytes than slower cell sites. The speed of data transfer is measured in bits per second (bps).

[0035] As the number of users increases, so does the amount of traffic generated in a given period of time. As a result, cell sites 122, 123, 130, 131, 141, 142 become increasingly congested. As a result, channels for cell sites 122, 123, 130, 131, 141, 142 continue to decrease. According to at least one embodiment, NMS 170 identifies current critical cells and anticipates when other cells will pose a problem that should be addressed.

[0036] FIG. 2 illustrates a key performance indicator (KPI) list 200 according to at least one embodiment.

[0037] In Figure 2, a KPI list 200 associated with a cell site is collected by a network management system. Multiple monitors collect KPIs for identifying capacity issues at the cell site. The KPI list 200 is used to check KPIs that affect the capacity of the device, and by analyzing the KPI data, it is possible to predict future usage and identify suspected future capacity violations.

[0038] The KPI list 200 shows downlink (DL) physical resource block (PRB) utilization parameters 210, an indicator of total traffic 212, radio resource control (RRC) connected users 214, identification of active user equipment (UE) 216, identification of Voice over Long Term Evolution (VoLTE) connected users 218, an indication of PRB usage 220 for guaranteed bit rate (GBR) traffic, and a measure of Internet Protocol (IP) DL throughput 222.

[0039] DL PRB Utilization 210 indicates the average PRB utilization per TTI (Transmission Time Interval) in the downlink direction. The utilization is defined by the ratio of used PRBs to available PRBs per TTI. DL PRB Utilization 210 is used to manage Quality of Service (QoS). Total Traffic 212 is a measure of the total amount of data messages received or transmitted over the communication channel. RRC Connected Users 214 is a measure of the total number of users connected to the RRC. Active UE 216 is the total number of active users currently connected to the network where data is being sent or communication is taking place. VoLTE Connected Users 218 is a measure of the number of users connected to Voice over Long-Term Evolution (VoLTE). PRB Usage for GBR Traffic 220 is a measure of the usage of radio resources of each cell for GBR traffic. IP DL Throughput 222 is a measure of the IP protocol specific DL throughput for the cell site.

[0040] The KPI list 200 is used to forecast device capacity to identify future issues in the network due to capacity violations. The capacity of any network is key to performance and if not properly managed, network problems will occur. A device according to at least one embodiment helps to predict capacity violations of network devices in a given window in the future, which helps to reduce the number of problems in the network due to capacity violations.

[0041] FIG. 3 illustrates an orientation and angle calculation 300 according to at least one embodiment.

[0042] In FIG. 3, a cell 310 is shown pointing in a particular direction. The azimuth angle is based on a beam width angle 312 of 30 degrees on either side of the center of the beam 314 for a total of 60 degrees. A predetermined distance, e.g., 300 meters (314), for a radius using a step size of 2 degrees. The step size means how much distance (in meters) should be plotted to draw the boundary. For example, to draw one circle, multiple data points should be kept as close as possible for better accuracy. If too large a step size, i.e., too large a distance, is used, the accuracy of forming the circle is lost. The predetermined distance is set so that the beam 314 captures everything up to 300 meters (316) within an azimuth angle of 60 degrees (312). If no cell site is found within the area 318, the region is expanded (e.g., the range of 300 meters (314) can be increased). Neighboring cell sites are determined for the area 318 defined by the azimuth angles 312 and 300 meters (316). 3, cell sites 320, 322, 324, and 326 are within area 318. Cells that are not within 300 meters 316 and are outside angle 312 are identified. For example, cell site 330, planned cell site 340, and new on-air cell site 350 are outside area 318.

[0043] 4 and 4C are a flowchart 400 of a method for identifying critical and non-critical cells according to at least one embodiment.

[0044] In Figures 4 and 4C, the process starts (S410) and the last six months of data for the KPIs is obtained from the KPI list (S414).

[0045] The data is separated into a first historical time window, e.g., Q1 (first 3 months of historical data), and a more recent historical time window, e.g., Q2 (last 3 months of historical data) (S418). The Q1 data and Q2 data are stored (S422).

[0046] The distance and azimuth angle calculation is used to determine the closest "last month on air" cell site within the area defined by a distance of 300 meters and an azimuth angle of 60 degrees (S426). In response to the distance and azimuth angle calculation being "yes" (S430), the "last month on air" cell site (S430) located within the area defined by a distance of 300 meters and an azimuth angle of 60 degrees is excluded (S434). The process then stops (S438).

[0047] In response to the distance and azimuth angle calculation being "no" (S442), identify whether the cell meets the criteria for a critical cell (S446). The KPIs are checked to determine whether the predefined KPIs meet the threshold criteria during a recent predefined time window, e.g., whether the DL PRB exceeds 70% during any 63 days out of any 90 days in Q2, as well as during a longer predefined historical time window, e.g., whether the DL PRB exceeds 70% during any 126 days out of any 180 days in Q1 and Q2. Those skilled in the art will appreciate that although months, days, and quarters are used to describe at least one embodiment herein, other time frames or windows may also be used.

[0048] In response to identifying cells that violate the condition (S450), these cells are listed as critical cell sites and considered for immediate expansion (S454). Critical cells are highly utilized and the cells are considered for capacity increase action and the process continues at A1 of FIG.

[0049] In response to the cell not meeting the critical cell criteria (S458), the data is smoothed (S462). The data is smoothed to remove seasonality and noise from the data. In at least one embodiment, a moving average technique is used to smooth the data. The moving average is calculated by summing all data points during a particular period and dividing the sum by the number of time periods. For anomalous days where data is not present for any reason, the previous valid value is considered for that day. For example, if the cell site is not emitting, the cell site is closed on that day, or for any reason data is not received from the device, a hyphen (-) value is received for that particular day. A 0 (zero) value is due to a slope change in the CM parameter, and the KPI data is downgraded and populated with a 0 (zero) value. For days with a 0 (zero) or hyphen (-) value, the last available non-zero value for those anomalous days is considered. Thus, days with either a "0" value or a "-" value have the last available data copied to their value for those days. The data is sampled by taking the moving average of the last 7 days. The sampling period remains the same for all cells, e.g., 1st to 7th of January.

[0050] A linear regression is then performed on the Q1 and Q2 data samples and the slope is obtained to identify trends (S466).

[0051] In at least one embodiment, the slope is calculated according to:

number

[0052]

number

number

[0053] The slope is based on the equation of a line, y=mx+c, generated by applying linear regression to the Q1 and Q2 data samples. This data set has non-critical cells to which a predictive model is applied to predict the data. In response to a positive slope (+m), the trend is considered to be increasing. In response to a negative slope (-m), the trend is considered to be decreasing. In response to the slope being stable (m=0), the trend is considered to be stable. For each value of m, the lookup table is updated. These cell sites are shortlisted as non-critical cells (S470) and the process continues to A2 of FIG. 7.

[0054] FIG. 5 is a flowchart 500 of a method for handling critical cells without prediction, according to at least one embodiment.

[0055] In Figure 5, the process continues from A1 of Figure 4C, where a list of critical cells is received (S510). These cells have already been identified as being critical cells and are highly utilized, so no predictive process is applied. The last six months of cell level data for the total traffic KPI is used to process the critical cells (S514). Linear regression is applied to the cell level data to identify trends associated with the critical cells (S516).

[0056] In Q2, it is determined whether the cell has a negative trend (S518). In response to the cell being determined not to have a negative trend in Q2 (S522), the cell is marked as a critical cell (S536), and the process proceeds to B1 of FIG. 8 (S560). In response to the cell being determined to have a negative trend in Q2 (S526), ​​it is determined whether the cell has a negative trend in Q1+Q2 (S530). In response to the cell being determined not to have a negative trend in Q1+Q2 (S534), the cell is marked as a critical cell (S536), and the process proceeds to B1 of FIG. 8 (S560). In response to the cell being determined to have a negative trend in Q1+Q2 (S544), the cell data is stored (S548), and an RCA analysis is performed to determine the cause of the negative trend in Q1+Q2 (S552). For example, it may be determined that a large number of users in the area served by the cell site have moved into the area and the trend is a long-term trend. Alternatively, it may be determined that a large number of users are moving into the area because an event has occurred and the trend is considered to be a temporary trend. After the event ends, the utilization of the cell site will decrease again and the trend will become negative, so the data is stored for later RCA analysis to determine why the trend remains negative. The process then ends (S554).

[0057] FIG. 6 illustrates a look-up table 600 for determining how to make a forecasting decision, according to at least one embodiment.

[0058] In FIG. 6, a first column is provided for Q1 trend 610. A second column is provided for Q2 trend 620. Next, columns are provided for identifying expected decision combinations 630. Columns are provided for identifying expected decision outcomes 650. In the first row, the Q1 trend 610 is increasing (X) 612. The Q2 trend is increasing (X) 622. The expected decision combination 630 is X-X 631. The expected decision outcome is a forecast 652 based on 6 months (Q1+Q2) data. The Q1 trend 610 can also be stable (Z) 614, decreasing (Y) 616, and data unavailable (W) 618. The Q2 trend 620 can also be stable (Z) 624, and decreasing (Y) 626. The forecast determination combinations 630 include X-Y 632, X-Z 633, Y-X 634, Y-Y 635, Y-Z 636, Z-X 637, Z-Y 638, Z-Z 639, W-X 640, W-Y 641, and W-Z 642. Additionally, the forecast determination results 640 include a forecast 654 based on 3-month (Q2) data.

[0059] FIG. 7 is a flowchart 700 of a method for handling non-critical cells using prediction, according to at least one embodiment.

[0060] In Figure 7, a list of non-critical cells is received (S710). Based on a look-up table (e.g., as shown in Figure 6), a forecast decision is made for either the last 3 months (Q2) or the last 6 months data (Q1+Q2) (S714). The last 6 months cell level data of total traffic KPI is used to process the critical cells (S718).

[0061] It is determined whether the cell has a negative trend in Q2 (S722). In response to the cell being determined not to have a negative trend in Q2 (S726), the process loops to perform predictions to determine the number of cells that will hit the target PRB threshold in the next 12 months (S760). The planned cell data and neighboring cell data are taken into account in predicting the outcome for the cell, thereby improving the prediction outcome and providing better accuracy. The prediction can reduce the number of problems due to network capacity, allowing better service to be provided to users. The process then proceeds to B2 in FIG. 8.

[0062] In response to the cell being determined to have a negative trend in Q2 (S730), it is determined whether the cell has a negative trend in Q1+Q2 (S734). In response to the cell being determined to not have a negative trend in Q1+Q2 (S736), the process loops to apply a prediction model to determine the number of cells that will hit the target PRB threshold in the next 12 months (S760). The planned cell data and neighboring cell data are taken into account in predicting the outcome for the cell, thereby improving the prediction outcome and providing better accuracy. The prediction can reduce the number of problems caused by network capacity, allowing better service to be provided to users. The process then proceeds to B2 in FIG. 8.

[0063] In at least one embodiment, a Seasonal AutoRegressive Integrated Moving Average (SARIMA) forecasting model is used to forecast cell KPIs and determine the dates when a cell violates a threshold (e.g., DL PRB exceeds 70%). The SARIMA model supports univariate time series data with a seasonal component. The SARIMA model adds three hyperparameters to specify the autoregressive (AR), differencing (I) and moving average (MA) of the seasonal component of the series, as well as an additional parameter for the period of seasonality. To configure SARIMA, the trend and seasonal components of the series are set. The trend components include p (trend autoregressive order), d (trend differencing order), and q (trend moving average order). There are four seasonal components that are not part of ARIMA that are configured for SARIMA. The four seasonal components are P (seasonal autoregressive order), D (seasonal differencing order), Q (seasonal moving average order), and m (number of time steps in a single seasonal period). In conjunction, the notation for the SARIMA model is designated as SARIMA(p,d,q)(P,D,Q)m. In at least one embodiment, the values ​​of the above parameters are p=0, d=1, q=0, P=1, D=1, Q=1, m=7, with m=7 for daily data suggesting a weekly seasonal cycle.

[0064] In response to the cell being determined to have a negative trend in Q1+Q2 (S742), the cell data is stored (S746) and an RCA analysis is performed (S750) to determine the cause of the negative trend in Q1+Q2 as to why the cell's trend is negative, after which the process ends (S754).

[0065] FIG. 8 is a flowchart 800 of a method for taking actions to address identified capacity issues, according to at least one embodiment.

[0066] In FIG. 8, a critical cell from B1 of FIG. 5 is received (S810). A prediction of non-critical cells meeting the target threshold in a given prediction time window is received from B2 of FIG. 7 (S820). The cells are predicted to hit the target threshold in a given window such as 0-3 months (S822), 3-6 months (S824), 6-9 months (S826), and 9-12 months (S828). A P1 priority cell (S830) is predicted to hit the target threshold in 0-3 months (S822). A P2 priority cell (S832) is predicted to hit the target threshold in 3-6 months (S824). A P3 priority cell (S834) is predicted to hit the target threshold in 6-9 months (S826). A P4 priority cell (S836) is predicted to hit the target threshold in 9-12 months (S828). Those skilled in the art will recognize that while at least one embodiment is described using predefined time windows of 0-3 months (S822), 3-6 months (S824), 6-9 months (S826), and 9-12 months (S828), other time frames may be used instead. In at least one embodiment, the predefined time windows are configurable. Furthermore, those skilled in the art will recognize that running the predictive model every day will not provide meaningful data because critical cells have already been identified. Thus, the predictive model is applied using a longer cycle, such as one month, to identify new critical cells with predicted high utilization in future windows.

[0067] The data about the critical and non-critical cells is then stored in a database (S840). A report is automatically generated based on the data (S850). Actions from the report are then taken (S860) to configure the network to address at least one of the current critical cells and / or cells having predicted capacity problems within a predetermined time window.

[0068] Those skilled in the art will appreciate that it is possible to perform actions from the report to configure the network to address the current critical cell, a cell with predicted capacity problems within a predefined time window, or both the current critical cell and a cell with predicted capacity problems within a predefined time window. Furthermore, "actions from the report are performed to configure the network to address at least one of the current critical cell, or a cell with predicted capacity problems within a predefined time window" is understood to mean that actions from the report are performed to configure the network to address the current critical cell, a cell with predicted capacity problems within a predefined time window, or both the current critical cell and a cell with predicted capacity problems within a predefined time window.

[0069] There are many ways to increase the capacity of a cell site. In at least one embodiment, actions to address capacity issues include expanding the capacity of cells by increasing one or more of resources, bandwidth, antennas, etc. in critical cells and / or cells with predicted capacity issues within a given time window, downgrading non-critical cells by reducing one or more of resources, bandwidth, antennas, etc., reducing the load on cells, installing new cells in close proximity to critical cells and / or cells with predicted capacity issues within a given time window, modifying routers and switches in cells, upgrading critical cells and / or cells with predicted capacity issues within a given time window from 3G to 4G and 5G, and downgrading critical cells and / or cells with predicted capacity issues within a given time window from 5G to 4G.

[0070] For example, LTE networks are significantly more spectrally efficient than 3G in carrying mobile traffic. LTE is capable of further improvements, e.g., LTE Advanced (LTE-A or 4G+). These improvements fall into three categories: 1) increasing the raw transmission throughput over the LTE radio link; 2) further enhancing the possibilities for spectrum reuse; and 3) more efficiently packing the offered traffic into the available transmission capacity.

[0071] LTE also implements higher-order multiple-input multiple-output (MIMO) implementations to provide increased traffic capacity. Some LTE deployments use 2×2 MIMO, whereby two antennas are placed at the base station and two antennas are placed at the user device. Due to the slight physical displacement of each transmit antenna from the other transmit antennas and each receive antenna from the other receive antennas, each sent and received signal will experience different multipath characteristics. By examining the four signals together, more of the originally coded information can be extracted. In addition, MIMO techniques may be used to send multiple simultaneous transmit streams between the base station and the user device.

[0072] 5G offers even more bandwidth or capacity than 4G. This is because 5G uses the available spectrum much more efficiently. While 4G uses a narrow slice of the available spectrum between 600 MHz and 2.5 GHz, 5G is split into three different bands. Each band has its own frequency range and speeds, and has different applications and use cases.

[0073] In at least one embodiment, in order to reduce cell size and therefore increase the number of cells serving a given area, it is possible to increase the amount of data carried by a cell by splitting or dividing the cell, which reduces congestion and interference by increasing the carrying capacity of the channel, increases network availability and reliability, and provides a higher degree of frequency reuse.

[0074] Splitting cells is implemented by deploying more radio towers / antennas and reducing the reach of each tower by reducing the radiated power of its radio transmissions. This allows the radio spectrum to be reused for multiple simultaneous transmissions within a geographic area. Thus, subdividing cells increases the amount of traffic that the spectrum can carry within an entire geographic area. Each cell may utilize its own base station, and reductions in antenna height and transmitter power may also be implemented.

[0075] Cell sectorization is another technique used to enhance capacity. In cell sectorization, each cell is subdivided into radial sectors with directional BS antennas to improve the performance of the system to combat interference caused by co-channels. A number of sectorized antennas are mounted on a single microwave tower located at the center of the cell, and a following number of antennas are installed to cover the 360-degree area of ​​the cell. During the process of cell sectorization, the number of cells that make up a particular cluster is reduced, and the distance separating the co-channels also becomes closer to each other. Thus, cell sectorization reduces the co-channel interference to enhance the capacity of the cellular system.

[0076] The above represent just a few examples of actions that may be taken to increase the capacity of a cell. Those skilled in the art will recognize that other methods, either existing or developed in the future, may be utilized without departing from the scope of the embodiments described herein. Operators may also change techniques or methods in response to a cell exhibiting a negative trend.

[0077] After a report is generated (S850) and actions are taken to address the capacity issue (S860), the process ends (S870). For example, a user can download a report on all cells that are currently highly utilized or will be highly utilized in the near future (within a given time window), and based on the report, the user can take actions such as planning and installing new sites in the area to improve network coverage. KPI data is also visible on a dashboard for monitoring purposes. The user can also perform other processes or use other tools based on the analysis provided in the report to determine the coverage in the area. Statistical analysis can be performed to gain deeper insight into the capacity utilization of cells in the network.

[0078] FIG. 9 illustrates a predetermined prediction window 900 according to at least one embodiment.

[0079] In FIG. 9, a prediction is made on historical data obtained from one or more of a first historical data window, e.g., Q1 910, and a second historical data window, e.g., Q2 920, for a current date 930, as described above. A timeline 932 shows a historical time frame 934 and a prediction time frame 936. A prediction is made based on one or more of the first historical data window 910 and the second historical data window 920. A prediction is made for predefined prediction windows, e.g., 0 months to 3 months 940, 3 months to 6 months 950, 6 months to 9 months 960, and 9 months to 12 months 970, for the current date 930. A 3 month window is used for illustrative purposes, but in at least one embodiment, the window length has a different length, e.g., 2 months, 4 months, etc.

[0080] FIG. 10 illustrates a capacity planning report 1000 according to at least one embodiment.

[0081] 10, capacity planning report 1000 includes a current critical cell identifier (ID) 1010. The current critical cell 1010 is identified as "highly utilized" (1012). Actions to be taken to address the current critical cell are listed in the report (1014).

[0082] The report also includes predicted critical cells 1020, which are shown as "Trend Up: Identify Time Frame for Capacity Expansion" (1022). The report lists actions to be taken (1024) to address cells with predicted capacity issues within a given time window (to increase the predicted capacity of critical cells (from non-critical cells) within a given time window).

[0083] 10 also shows cells with predicted capacity reductions 1030. Cells with capacity reductions 1030 are shown as "Trend Reduction: Identify time frame for contraction" (1032). Actions to be taken to address cells with predicted capacity issues within a given time window (to address the reduction in capacity of the cell within a given time window) are listed in the report (1034).

[0084] The location of demand can change over time, sometimes unpredictably. Other technologies, such as full fiber broadband and Wi-Fi, continue to advance and serve some of the same data use cases that can affect demand over time.

[0085] FIG. 11 is a flowchart 1100 of a method for anticipating capacity violations in a mobile network according to at least one embodiment.

[0086] In Figure 11, the method starts (S1110) and data relating to key performance indicators is obtained (S1114). Referring to Figure 2, multiple monitors collect KPIs for identifying capacity issues at a cell site. A KPI list 200 is used to check KPIs that affect the capacity of the device, and by analyzing the KPI data, it is possible to predict future usage and identify suspected future capacity violations. The KPI list 200 shows downlink (DL) physical resource block (PRB) utilization parameters 210, an indicator of total traffic 212, radio resource control (RRC) connected users 214, identification of active user equipment (UE) 216, identification of voice over long term evolution (VoLTE) connected users 218, an indication of PRB usage 220 for guaranteed bit rate (GBR) traffic, and a measurement of Internet Protocol (IP) DL throughput 222.

[0087] Sites to be excluded are identified (S1118). With reference to Figure 4, for example, in response to the distance and azimuth angle calculation being "yes" (S430), a "last month on air" cell site (S430) located within an area defined by a distance of 300 meters and an azimuth angle of 60 degrees is excluded (S434). Critical cells having a predefined KPI above a threshold are determined (S1122).

[0088] For the remaining non-critical cells, the data is processed to smooth the data (S1126). Referring to FIG. 4C, to smooth the data, days with 0 (zero) or hyphen (-) values ​​are removed and the last available non-zero values ​​of those anomalous days are considered. Thus, days with either "0" or "-" values ​​have the last available data copied to those days' values. The data is sampled by taking the moving average of the last 7 days. The sampling period remains the same for all cells, for example, January 1st to 7th.

[0089] A linear regression is applied (S1130) to determine the cell capacity trend of the non-critical cells. Referring to FIG. 4C, the slope is based on the equation of a line, y=mx+c, generated by applying linear regression to the Q1 and Q2 data samples. This data set has non-critical cells to which a predictive model is applied to forecast the data. In response to a positive slope (+m), the trend is considered to be increasing. In response to a negative slope (-m), the trend is considered to be decreasing. In response to the slope being stable (m=0), the trend is considered to be stable. For each value of m, the lookup table is updated. Those skilled in the art will appreciate that although months, days, and quarters are used herein to describe at least one embodiment, other time frames or windows may also be used.

[0090] The prediction model is applied to predict future capacity problems using the latest history time window or the last two latest history time windows (S1134). Referring to Figure 7, the planned cell data and neighboring cell data are taken into account in predicting the outcome for the cell, thereby improving the prediction result and providing better accuracy. The prediction can reduce the number of problems caused by network capacity and provide better service to users.

[0091] A predefined time window is determined (S1138) to anticipate further capacity issues. For example, referring to FIG. 7, a prediction is performed to determine the number of cells that will hit the target TH within a predefined time window, e.g., the PRB threshold in the next 12 months (S760). Referring to FIG. 8, cells are predicted to hit the target threshold in predefined windows such as 0 months to 3 months (S822), 3 months to 6 months (S824), 6 months to 9 months (S826), and 9 months to 12 months (S828). Those skilled in the art will recognize that although at least one embodiment is described using predefined time windows of 0 months to 3 months (S822), 3 months to 6 months (S824), 6 months to 9 months (S826), and 9 months to 12 months (S828), additional time frames, fewer time frames, or different length time frames may be used instead.

[0092] A report is generated (S1142) for the critical cells and non-critical cells that are predicted to have capacity problems in the given time window. Referring to FIG. 10, the capacity planning report 1000 includes a current critical cell identifier (ID) 1010. The current critical cell 1010 is identified as being highly utilized 1012. Actions 1014 to address the current critical cell are listed. The report also includes a predicted critical cell 1020. The predicted critical cell 1020 is shown as "trend up: identify time frame for capacity expansion" 1022. Actions 1024 to take to increase capacity for cells that have predicted capacity problems in the given time window are shown. FIG. 10 also shows cells with predicted capacity reduction 1030. The cells with capacity reduction 1030 are shown as "trend down: identify time frame for contraction" 1032. Actions 1024 to take to address cells that have predicted capacity problems in the given time window are shown.

[0093] Actions from the report are taken to configure the network to address at least one of the current critical cells and / or cells having predicted problems with a predefined time window (S1146). Referring to Figure 8, based on the generated report, actions are taken to configure the network to address the current critical cells and / or cells having predicted capacity problems within a predefined time window (S860).

[0094] Those skilled in the art will appreciate that it is possible to perform actions from the report to configure the network to address the current critical cell, a cell with predicted capacity problems within a predefined time window, or both the current critical cell and a cell with predicted capacity problems within a predefined time window. Furthermore, "actions from the report are performed to configure the network to address at least one of the current critical cell, or a cell with predicted capacity problems within a predefined time window" is understood to mean that actions from the report are performed to configure the network to address the current critical cell, a cell with predicted capacity problems within a predefined time window, or both the current critical cell and a cell with predicted capacity problems within a predefined time window.

[0095] Furthermore, as discussed above, there are many ways to increase the capacity of a cell site. In at least one embodiment, actions to address capacity issues include expanding the capacity of the cells by increasing at least one of resources and bandwidth in the critical cells and cells with predicted capacity issues within a predetermined time window, downgrading non-critical cells by reducing at least one of resources and bandwidth, reducing the load on the cells, installing new cells in close proximity to the critical cells and cells with predicted capacity issues within a predetermined time window, modifying routers and switches in the cells, upgrading the critical cells and cells with predicted capacity issues within a predetermined time window from 3G to 4G and 5G, and downgrading the critical cells and cells with predicted capacity issues within a predetermined time window from 5G to 4G.

[0096] The process then ends (S1150).

[0097] At least one embodiment for predicting capacity violations in a mobile network includes accessing a KPI database to obtain key performance indicator (KPI) data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, where the critical cells exhibit high utilization rates that affect performance by the critical cells and the non-critical cells do not exhibit high utilization rates; for the non-critical cells, applying a predictive model to identify at least one predefined prediction time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the predictive model that identifies at least one action to be taken to configure the mobile network to address at least one of the capacity problems of the critical cells or the non-critical cells having a predicted capacity problem within one of the predefined prediction time windows; and performing the at least one action to configure the mobile network to address at least one of the capacity problems of the critical cells or the non-critical cells having a predicted capacity problem within one of the predefined prediction time windows.

[0098] FIG. 12 is a high-level functional block diagram of a processor-based system 1200 according to at least one embodiment.

[0099] In at least one embodiment, the processing circuit 1200 links at least one subsequent correlated alarm to the primary alarm. The processing circuit 1200 implements a method for predicting capacity violations in a mobile network using a processor 1202. The processing circuit 1200 also includes a non-transitory computer-readable storage medium 1204 used to implement the method for predicting capacity violations in a mobile network. The storage medium 1204 is encoded with, i.e. stores, instructions 1206, i.e. computer program code, that are executed by the processor 1202, among other things, to cause the processor 1202 to perform operations for predicting capacity violations in a mobile network. The execution of the instructions 1206 by the processor 1202 represents (at least in part) an application that implements at least a portion of the methodology described herein (hereinafter the described process and / or methodology) in accordance with one or more embodiments. In at least one embodiment, the processing circuit 1200 is a server, such as a cloud server, that accesses a KPI database 1226.

[0100] The processor 1202 is electrically coupled to the computer-readable storage medium 1204 via a bus 1208. The processor 1202 is electrically coupled to an input / output (I / O) interface 1210 by the bus 1208. A network interface 1212 is also electrically connected to the processor 1202 via the bus 1208. The network interface 1212 is connected to a network 1214, thereby connecting the processor 1202 and the computer-readable storage medium 1204 to external elements via the network 1214. The processor 1202 is configured to execute the instructions 1206 encoded in the computer-readable storage medium 1204, enabling the processing circuit 1200 to execute at least a portion of a process and / or method. In one or more embodiments, the processor 1202 is a central processing unit (CPU), a multiprocessor, a distributed processing system, an application specific integrated circuit (ASIC), and / or any other suitable processing unit.

[0101] The processing circuit 1200 includes an I / O interface 1210. The I / O interface 1210 is coupled to external circuitry. In one or more embodiments, the I / O interface 1210 includes a keyboard, a keypad, a mouse, a trackball, a trackpad, a touch screen, and / or cursor direction keys for communicating information and commands to the processor 1202.

[0102] The processing circuit 1200 also includes a network interface 1212 coupled to the processor 1202. The network interface 1212 enables the processing circuit 1200 to communicate with a network 1214 to which one or more other computer systems are connected. The network interface 1212 includes a wireless network interface, such as Bluetooth, Wi-Fi, Worldwide Interoperability for Microwave Access (WiMAX), General Packet Radio Service (GPRS), or Wideband Code Division Multiple Access (WCDMA), or a wired network interface, such as Ethernet, Universal Serial Bus (USB), or Institute of Electrical and Electronics Engineers (IEEE) 1264.

[0103] The processing circuit 1200 is configured to receive information via an I / O interface 1210. The information received via the I / O interface 1210 includes one or more of instructions, data, design rules, libraries of cells, and / or other parameters for processing by the processor 1202. The information is transferred to the processor 1202 via the bus 1208. The processing circuit 1200 is configured to receive information related to a user interface (UI) 1222 via the I / O interface 1210. The information is stored in the computer-readable medium 1204 as a UI 1222. The UI 1222 is presented on a display device 1224 to forecast capacity violations in the mobile network. KPI data is also presented in the UI 1222 in a dashboard to monitor cell performance / capacity. In at least one embodiment, a report identifying at least one action to be taken to address current critical cells or cells having forecasted capacity issues within one predefined time window is automatically displayed in the UI 1222 by the display device 1224. Those skilled in the art will appreciate that it is possible to perform actions from the report to configure the network to address the current critical cell, a cell with predicted capacity problems within a predefined time window, or both the current critical cell and a cell with predicted capacity problems within a predefined time window. Furthermore, "actions from the report are performed to configure the network to address at least one of the current critical cell, or a cell with predicted capacity problems within a predefined time window" is understood to mean that actions from the report are performed to configure the network to address the current critical cell, a cell with predicted capacity problems within a predefined time window, or both the current critical cell and a cell with predicted capacity problems within a predefined time window.

[0104] One or more non-transitory computer readable storage media 1204 that store instructions (in compressed or uncompressed form) used to program a computer, processor, or other electronic device to perform a process or method described herein in one or more embodiments. The one or more non-transitory computer readable storage media 1204 include one or more of electronic storage media, magnetic storage media, optical storage media, quantum storage media, etc. For example, computer readable storage media include, but are not limited to, hard drives, floppy diskettes, optical disks, read only memory (ROM), random access memory (RAM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic or optical cards, solid state memory devices, or other types of physical media suitable for storing electronic instructions. In one or more embodiments using optical discs, the one or more non-transitory computer-readable storage media 1204 include a compact disc read only memory (CD-ROM), a compact disc read / write (CD-R / W), and / or a digital video disc (DVD).

[0105] In one or more embodiments, the storage medium 1204 stores computer program code 1206 configured to cause the processing circuit 1200 to execute at least a portion of a process and / or method for providing a subsequent correlated alarm list. In one or more embodiments, the storage medium 1204 also stores information, such as algorithms, that facilitate execution of at least a portion of a process and / or method for predicting capacity violations in a mobile network. Thus, in at least one embodiment, the processor circuit 1200 executes a method for predicting capacity violations in a mobile network.

[0106] The process includes accessing a KPI database to obtain key performance indicator (KPI) data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, where the critical cells exhibit high utilization rates that affect performance by the critical cells and the non-critical cells do not exhibit high utilization rates; applying a predictive model for the non-critical cells to identify at least one predefined predicted time window associated with a capacity problem associated with at least one of the non-critical cells; and performing actions to configure the mobile network to address the capacity problem of the critical cell and the capacity problem of the non-critical cells that have a predicted capacity problem within one of the predefined predicted time windows.

[0107] The process for predicting capacity violations in mobile networks has the advantages of identifying cells that will be important in the near future, reducing capacity and network failures due to current capacity loads, improving network performance, proactively managing highly utilized cells, and identifying potential risks in the network. Predicting future network capacity violations helps to plan network expansion to avoid network failures due to capacity violations.

[0108] In at least one embodiment, separate instances of these programs are executed or distributed on any number of separate computer systems. Thus, although particular steps are described as being executed by particular devices, software programs, processes, or entities, this is not necessarily the case. Various alternative implementations will be appreciated by those skilled in the art.

[0109] Moreover, those skilled in the art will readily recognize that the above-described techniques can be utilized in a variety of devices, environments, and contexts. Although the embodiments have been described in language specific to structural features or methodological acts, the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described. Rather, the specific features and acts are disclosed as example forms of implementing the claims.

Claims

1. 1. A method for predicting capacity violations in a mobile network, comprising: accessing a key performance indicator (KPI) database to obtain KPI data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, the critical cells exhibiting high utilization rates that impact performance due to the critical cells, and the non-critical cells not exhibiting high utilization rates; applying a predictive model to identify, for the non-critical cells, at least one predetermined forecasted time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the prediction model that identifies at least one action to be taken to configure the mobile network to address at least one of a capacity problem of the critical cell or a capacity problem of the non-critical cell having a predicted capacity problem within one of the predetermined prediction time windows; performing the at least one action to configure the mobile network to address the at least one of a capacity problem of the critical cell or a capacity problem of the non-critical cell having a predicted capacity problem within one of the predetermined prediction time windows; A method comprising:

2. 2. The method of claim 1, wherein the accessing the KPI database to obtain KPI data related to capacity of the cells in the mobile network further comprises obtaining KPI data for a first historical time window and a most recent historical time window, the first historical time window occurring immediately before the most recent historical time window.

3. identifying the critical cells determining cells in the predetermined area that are not new on-air cells; identifying, based on the KPI data, the cells in the predetermined area that exceed a predetermined performance threshold as the critical cells, rather than new on-air cells; The method of claim 1 , comprising:

4. identifying the cells in the predetermined area that exceed the predetermined performance threshold as the critical cells based on the KPI data; Determining, for a first quarter, an average downlink (DL) physical resource block (PRB) utilization rate that exceeds 63 days (70%) out of 90 days in the first quarter, or, for a second quarterly period, determining, for 126 days (70%) out of 180 days in the second quarterly period. The method of claim 3 , comprising:

5. identifying the non-critical cells determining cells in the predetermined area that are not new on-air cells; Identifying the cells in the predetermined area that are not new on-air cells and do not exceed a predetermined performance threshold as the non-critical cells based on the KPI data; filling the invalid data values ​​by copying the last available valid data for a day into the invalid data values ​​to generate adjusted data for the non-critical cells; calculating a moving average of the adjusted data to generate averaged data for the non-critical cells; applying a linear regression to the averaged data for the non-critical cells to identify trends associated with the averaged data; The method of claim 1 , comprising:

6. applying a linear regression to the averaged data to identify the trends associated with the average data for the non-critical cells; determining whether the non-critical cells have a negative trend associated with the averaged data for the non-critical cells during a previous quarter; determining whether the non-critical cell has the negative trend associated with the averaged data for the non-critical cell during the immediately preceding two quarters in response to determining that the non-critical cell has the negative trend associated with the averaged data for the non-critical cell during the immediately preceding quarter; in response to determining that the non-critical cell has a negative trend associated with the averaged data for the non-critical cell during the immediately preceding quarter and between the immediately preceding two quarters, storing the data associated with the non-critical cell and performing a root cause analysis using the data to identify reasons for the negative trend during the immediately preceding quarter and between the immediately preceding two quarters. Further comprising: responsive to determining that the non-critical cells have not had the negative trend during the immediately preceding quarter and during the immediately preceding two quarters, the forecast model is applied to the non-critical cells to identify at least one predetermined forecast time window associated with a capacity problem associated with at least one of the non-critical cells. The method according to claim 5.

7. applying the predictive model to identify the at least one predetermined forecast time window associated with the capacity problem associated with the at least one of the non-critical cells; applying a Seasonal Autoregressive Integrated Moving Average (SARIMA) forecasting model to the at least one of the non-critical cells; identifying a first predetermined forecast window for the non-critical cell's capacity problem expected to occur in 0 to 3 months, a second predetermined forecast window for the non-critical cell's capacity problem expected to occur in 3 to 6 months, a third predetermined forecast window for the non-critical cell's capacity problem expected to occur in 6 to 9 months, and a fourth predetermined forecast window for the non-critical cell's capacity problem expected to occur in 9 to 12 months based on applying the SARIMA forecast model to the at least one of the non-critical cells; applying a first priority to the non-critical cells that are expected to have capacity problems in 0 to 3 months, a second priority to the non-critical cells that are expected to have capacity problems in 3 to 6 months, a third priority to the non-critical cells that are expected to have capacity problems in 6 to 9 months, and a fourth priority to the non-critical cells that are expected to have capacity problems in 9 to 12 months; The method of claim 1 , comprising:

8. identifying the critical cells determining whether the critical cell has a negative trend associated with the averaged data for the critical cell during a previous quarter; marking the critical cell in response to determining whether the critical cell has the negative trend associated with the averaged data for the critical cell during the two immediately preceding quarters; in response to determining that the critical cell has a negative trend associated with the averaged data for the critical cell during the immediately preceding quarter and between the immediately preceding two quarters, storing cell data associated with the critical cell and performing a root cause analysis to identify reasons for the negative trend during the immediately preceding quarter and between the immediately preceding two quarters; marking the critical cell in response to determining that the critical cell has the negative trend associated with the averaged data for the critical cell during the immediately preceding quarter and between the immediately preceding two quarters; The method of claim 1 , comprising:

9. A device for predicting capacity violations in a mobile network, comprising: a memory storing computer readable instructions; Executing the computer readable instructions, accessing a key performance indicator (KPI) database to obtain KPI data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, the critical cells exhibiting high utilization rates that impact performance due to the critical cells, and the non-critical cells not exhibiting high utilization rates; applying a predictive model to identify, for the non-critical cells, at least one predetermined forecasted time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the prediction model that identifies actions to be taken to configure the mobile network to address at least one of a capacity problem of the critical cell or a capacity problem of the non-critical cell having a predicted capacity problem within one of the predetermined prediction time windows; a processor configured to: A device comprising:

10. 10. The device of claim 9, wherein the processor accesses the KPI database to obtain KPI data related to the capacity of the cell in the mobile network by obtaining KPI data for a first historical time window and a most recent historical time window, the first historical time window occurring immediately before the most recent historical time window.

11. The processor, determining cells in the predetermined area that are not new on-air cells; identifying, based on the KPI data, the cells in the predetermined area that exceed a predetermined performance threshold as the critical cells, rather than new on-air cells; identifying the critical cell by 10. The device of claim 9, wherein identifying the cells in the predetermined area that exceed the predetermined performance threshold as the critical cells based on the KPI data includes determining, for a first quarter, an average downlink (DL) physical resource block (PRB) utilization rate that exceeds 63 days (70%) of a 90-day period in the first quarter, or, for a second quarterly period, an average downlink (DL) physical resource block (PRB) utilization rate that exceeds 126 days (70%) of a 180-day period in the second quarterly period.

12. The processor, determining cells in the predetermined area that are not new on-air cells; Identifying the cells in the predetermined area that are not new on-air cells and do not exceed a predetermined performance threshold as the non-critical cells based on the KPI data; filling the invalid data values ​​by copying the last available valid data for a day into the invalid data values ​​to generate adjusted data for the non-critical cells; calculating a moving average of the adjusted data to generate averaged data for the non-critical cells; applying a linear regression to the averaged data for the non-critical cells to identify trends associated with the averaged data; The device of claim 9 , wherein the non-critical cells are identified by:

13. The processor, applying a Seasonal Autoregressive Integrated Moving Average (SARIMA) forecasting model to the at least one of the non-critical cells; identifying a first predetermined forecast window for the non-critical cell's capacity problem expected to occur in 0 to 3 months, a second predetermined forecast window for the non-critical cell's capacity problem expected to occur in 3 to 6 months, a third predetermined forecast window for the non-critical cell's capacity problem expected to occur in 6 to 9 months, and a fourth predetermined forecast window for the non-critical cell's capacity problem expected to occur in 9 to 12 months based on applying the SARIMA forecast model to the at least one non-critical cell; applying a first priority to the non-critical cells that are expected to have capacity problems in 0 to 3 months, a second priority to the non-critical cells that are expected to have capacity problems in 3 to 6 months, a third priority to the non-critical cells that are expected to have capacity problems in 6 to 9 months, and a fourth priority to the non-critical cells that are expected to have capacity problems in 9 to 12 months; 10. The device of claim 9, further comprising: a predictive model for detecting a capacity problem associated with the at least one of the non-critical cells, the predictive model being configured to predict a capacity problem associated with the at least one of the non-critical cells by:

14. The processor, determining whether the critical cell has a negative trend associated with the averaged data for the critical cell during a previous quarter; marking the critical cell in response to determining whether the critical cell has the negative trend associated with the averaged data for the critical cell during the two immediately preceding quarters; in response to determining that the critical cell has a negative trend associated with the averaged data for the critical cell during the immediately preceding quarter and between the immediately preceding two quarters, storing cell data associated with the critical cell and performing a root cause analysis to identify reasons for the negative trend during the immediately preceding quarter and between the immediately preceding two quarters; marking the critical cell in response to determining that the critical cell has the negative trend associated with the averaged data for the critical cell during the immediately preceding quarter and between the immediately preceding two quarters; The device of claim 9 , wherein the critical cell is identified by:

15. A non-transitory computer readable medium having computer readable instructions stored thereon that, when executed by a processor, cause the processor to: accessing a key performance indicator (KPI) database to obtain KPI data related to capacity of cells in the mobile network; identifying critical cells and non-critical cells based on the KPI data, the critical cells exhibiting high utilization rates that impact performance due to the critical cells, and the non-critical cells not exhibiting high utilization rates; applying a predictive model to identify, for the non-critical cells, at least one predetermined forecasted time window associated with a capacity problem associated with at least one of the non-critical cells; generating a report based on applying the prediction model that identifies at least one action to be taken to configure the mobile network to address at least one of a capacity problem of the critical cell or a capacity problem of the non-critical cell having a predicted capacity problem within one of the predetermined prediction time windows; performing the at least one action to configure the mobile network to address the at least one of a capacity problem of the critical cell or a capacity problem of the non-critical cell having a predicted capacity problem within one of the predetermined prediction time windows; A non-transitory computer-readable medium for causing operations to be performed, including:

16. 16. The non-transitory computer-readable medium of claim 15, wherein the accessing the KPI database to obtain KPI data related to capacity of the cell in the mobile network further comprises obtaining KPI data for a first historical time window and a most recent historical time window, the first historical time window occurring immediately before the most recent historical time window.

17. identifying the critical cells determining cells in the predetermined area that are not new on-air cells; identifying, based on the KPI data, the cells in the predetermined area that exceed a predetermined performance threshold as the critical cells, rather than new on-air cells; 20. The non-transitory computer readable medium of claim 15, comprising:

18. identifying the non-critical cells determining cells in the predetermined area that are not new on-air cells; Identifying the cells in the predetermined area that are not new on-air cells and do not exceed a predetermined performance threshold as the non-critical cells based on the KPI data; filling the invalid data values ​​by copying the last available valid data for a day into the invalid data values ​​to generate adjusted data for the non-critical cells; calculating a moving average of the adjusted data to generate averaged data for the non-critical cells; applying a linear regression to the averaged data for the non-critical cells to identify trends associated with the averaged data; 20. The non-transitory computer readable medium of claim 15, comprising:

19. applying the predictive model to identify the at least one predetermined forecast time window associated with the capacity problem associated with the at least one of the non-critical cells; applying a Seasonal Autoregressive Integrated Moving Average (SARIMA) forecasting model to the at least one of the non-critical cells; identifying a first predetermined forecast window for the non-critical cell's capacity problem expected to occur in 0 to 3 months, a second predetermined forecast window for the non-critical cell's capacity problem expected to occur in 3 to 6 months, a third predetermined forecast window for the non-critical cell's capacity problem expected to occur in 6 to 9 months, and a fourth predetermined forecast window for the non-critical cell's capacity problem expected to occur in 9 to 12 months based on applying the SARIMA forecast model to the at least one of the non-critical cells; applying a first priority to the non-critical cells that are expected to have capacity problems in 0 to 3 months, a second priority to the non-critical cells that are expected to have capacity problems in 3 to 6 months, a third priority to the non-critical cells that are expected to have capacity problems in 6 to 9 months, and a fourth priority to the non-critical cells that are expected to have capacity problems in 9 to 12 months; 20. The non-transitory computer readable medium of claim 15, comprising:

20. identifying the critical cells determining whether the critical cell has a negative trend associated with the averaged data for the critical cell during a previous quarter; marking the critical cell in response to determining whether the critical cell has the negative trend associated with the averaged data for the critical cell during the two immediately preceding quarters; in response to determining that the critical cell has a negative trend associated with the averaged data for the critical cell during the immediately preceding quarter and between the immediately preceding two quarters, storing cell data associated with the critical cell and performing a root cause analysis to identify reasons for the negative trend during the immediately preceding quarter and between the immediately preceding two quarters; marking the critical cell in response to determining that the critical cell has the negative trend associated with the averaged data for the critical cell during the immediately preceding quarter and between the immediately preceding two quarters; 20. The non-transitory computer readable medium of claim 15, comprising:

Images