Method and apparatus for handling network parameter recommendations

Abstract

Embodiments of the present disclosure provide a method (200) for handling network parameter recommendations from a plurality of applications (106a-106n) being executed within an automation platform (102). The method is performed by a network node (114) in communication with the automation platform (102). The method comprises receiving, from a first application, a network parameter recommendation comprising at least one recommended network parameter value for configuring at least one network parameter of the network node. The method comprises initiating configuration of the recommended at least one network parameter value when it has been determined to configure the at least one network parameter using the recommended at least one network parameter value and requesting the first application to recommend at least one alternative network parameter value when it has been determined not to configure the at least one network parameter using the recommended at least one network parameter value.

Description

[0001] METHOD AND APPARATUS FOR HANDLING NETWORK PARAMETER RECOMMENDATIONS

[0002] TECHNICAL FIELD

[0003] The present disclosure relates to wireless communication. More particularly, it relates to method, network node, and computer program products for handling network parameter recommendations from a plurality of applications executed in an automation platform.

[0004] BACKGROUND

[0005] Wireless communication systems are widely deployed to provide various types of communication content such as data, voice, video, and so on. A wireless communication system comprises a Radio Access Network, RAN, for connecting User Equipments, UEs, to a Core Network, CN. Traditionally, hardware and / or software of the particular RAN is vendor specific.

[0006] In particular, Open RAN, O-RAN, architecture enables multiple vendors to provide hardware and / or software to the RAN comprising one or more network nodes. Further, the O-RAN system supports two types of automation applications, Apps, namely, xApps and rApps. A rApp can be executed on a non-real time RAN Controller, RIC, within an automation platform to provide network parameter recommendations for configuring network parameters of the one or more network nodes in the RAN. Thus, the rApp may act as autonomous agents interacting with the one or more network nodes.

[0007] Further, the automation platform enables the rApp to obtain subscription for one or more topics with read and / or write access to acquire information about the one or more network nodes in the RAN and / or to provide network parameter recommendations for configuring the network parameters of the one or more network nodes. The network parameter recommendations may comprise network parameter values recommended for configuring the network parameters of the one or more network nodes. For example, if the rApp obtains a subscription for a topic like energy consumption (EnergyMeter.pmConsumedEnergy), the rApp receives updates (i.e., request to read) for each network node when there is an update on a power consumption of that network node. Similarly, if the rApp obtains a subscription for a topic like antenna tilt configuration, the rApp sends updates (i.e., request to write) to the antenna tilt for a group of cells, for example, for down tilt of 1 degree. SUMMARY

[0008] In the current wireless communication system, the network node may configure the network parameters using the recommended network parameter values (provided by the rApps) on a first-in-first-out, FIFO, basis or at least without considering any quality measures associated with the recommended network parameter values. This may potentially lead to conflicts between the network parameters recommended by different rApps, when the rApps optimize for different goals and use the same network parameters to achieve the different goals. As a result, there may be performance degradation of the network node when different network parameters are required for the achievement of each goal.

[0009] For example, a rApp intending to optimize a first Key Performance Indicator, KPI, metric associated with the network parameter can change the network parameter by inadvertently affecting / impacting other KPI metrics optimized by other rApps due to conflict / overlap in their recommended network parameter values. This may result in the performance degradation of the network node.

[0010] Consequently, there is a need for an improved method and arrangement for handling network parameter recommendations from a plurality of applications, which alleviates at least some of the above-cited problems.

[0011] It is therefore an object of the present disclosure to provide a method, a network node, and a computer program product for handling network parameter recommendations, to mitigate, alleviate, or eliminate all or at least some of the above-discussed drawbacks of presently known solutions.

[0012] This and other objects are achieved by means of a method, a network node, and a computer program product as defined in the appended claims. The term exemplary is in the present context to be understood as serving as an instance, example or illustration.

[0013] According to a first aspect of the present disclosure, a method for handling network parameter recommendations from a plurality of applications is provided. The plurality of applications are executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automation platform. The method is performed by a network node in communication with the automation platform. The method comprises receiving a network parameter recommendation from a first application among the plurality of applications. The network parameter recommendation comprising at least one recommended network parameter value for configuring at least one network parameter of the network node. The method comprises determining whether or not to configure the at least one network parameter using the recommended at least one network parameter value. The method comprises initiating configuration of the at least one recommended network parameter value when it has been determined to configure the at least one network parameter using the recommended at least one network parameter value. The method comprises requesting the first application to recommend at least one alternative network parameter value when it has been determined not to configure the at least one network parameter using the recommended at least one network parameter value.

[0014] In some embodiments, the step of determining whether or not to configure the at least one network parameter using the recommended at least one network parameter value comprises identifying a plurality of Key Performance Indicator, KPI, metrics associated with the at least one network parameter for which the at least one network parameter value is recommended. The method comprises determining at least one KPI metric from the plurality of KPI metrics of the at least one network parameter targeted by the recommended at least one network parameter value. The method comprises determining whether there exists any conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value. When it has been determined that there is not conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, the method comprises determining to configure the at least one network parameter using the recommended at least one network parameter value. When it has been determined that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, the method comprises determining not to configure the at least one network parameter using the recommended at least one network parameter value.

[0015] In some embodiments, the step of detecting whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value comprises selecting, from lookup tables, LUTs, built for a number of possible states of the network node, a LUT for a current state of the network node for which the at least one network parameter value is recommended. The method comprises determining, using the selected LUT, whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value.

[0016] In some embodiments, the LUT built for the state of the network node comprises an association value for each combination of a network parameter and a respective KPI metric, wherein the association value indicates an association of the KPI metric with the respective network parameter.

[0017] In some embodiments, the step of selecting the LUT for the current state of the network node is preceded by the following steps: identifying the number of possible states of the network node and building the LUTs for the number of possible states of the network node. Each possible state represents characteristics of the network node in terms of a plurality of network parameters and KPI metrics associated with each of the plurality of network parameters.

[0018] In some embodiments, the step of identifying the number of possible states of the network node comprises determining a maximum possible number of states of the network node by evaluating each of the plurality of network parameters of the network node and respective KPI metrics using a first learning model. The method comprises building the LUT for each state of the determined maximum possible number of states. The method comprises grouping the states of the network node when the LUTs associated with said states of the network node differ from each other by a pre-defined value, wherein the grouped states are identified as the number of possible states of the network node.

[0019] In some embodiments, the step of building the LUT for a state of the identified number of possible states of the network node comprises obtaining historical data collected for a predetermined time interval, said historical data comprising a plurality of network parameter values recommended by the plurality of applications, a state of the network node associated with at least one of the recommended plurality of network parameter values, at least one network parameter configured using at least one of the plurality of network parameter values, and the KPI metrics resultant from the configuration of said at least one network parameter. The method comprises determining the association value for each combination of one of the plurality of network parameters and one of respective KPI metrics, by evaluating the historical data using a second learning model.

[0020] In some embodiments, the second learning model evaluates the historical data to determine the association value based on at least one of: identification of at least one feature as an importance from the obtained historical data; interpolation between the states of the network node, when the states have been arranged in an order; similarity measures between the states of the network node, when the states have not been arranged in the order; and an input received from a radio domain expert.

[0021] In some embodiments, the step of selecting the LUT for the current state of the network node for which the at least one network parameter value is recommended comprises: detecting, using a distance metric, whether any of the identified number of states of the network node matches with the current state of the network node for which the at least one network parameter value is recommended. When it has been determined that one of the identified number of states of the network node matches with the current state of the network node the method comprises selecting the LUT corresponding to said state as the LUT for the current state of the network node.

[0022] In some embodiments, the step of determining, using the selected LUT, whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value comprises: determining, using the selected LUT, an weighted sum of association value of the recommended at least one network parameter value with respect to each of the other KPI metrics different from the targeted KPI metric of the at least one network parameter. The method comprises comparing the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics with a threshold factor. The method comprises determining that there exists no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value when the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics is smaller than the threshold factor. The method comprises determining that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value when the weighted sum of association value of the at least one network parameter value is greater than the threshold factor.

[0023] In some embodiments, the step of determining, using the selected LUT, the weighted sum of association value of the recommended at least one parameter value with respect to each of the other KPI metrics different from the targeted KPI metric of the at least one network parameter comprises: identifying from the selected LUT a priority and an association value associated with each of the other KPI metrics with respect to the at least one network parameter. The method comprises determining weighted results for the other KPI metrics based on the priority and the association value of each of the other KPI metrics. The method comprises summing up the weighted results to determine the weighted sum of association value of the recommended at least one parameter value.

[0024] In some embodiments, the method further comprises a step of tuning the threshold factor. The step comprises evaluating the historical data using a third learning model to determine the at least one network parameter configured with a minimum configuration time and a minimum impact on other KPI metrics different from the targeted KPI metric of the at least one network parameter. The method comprises tuning the threshold factor based on the evaluation.

[0025] In some embodiments, the method further comprises a step of updating the LUT. The step comprises receiving update information related to at least one application and updating the LUT in accordance with the received update information. The update information indicates at least one of: on-boarding of a new application, and a subscription of the at least one application targeting new KPI metrics.

[0026] In some embodiments, the step of determining whether or not to configure the at least one network parameter using the recommended at least one network parameter value comprises determining an weighted sum of association value of the recommended at least one network parameter value using a LUT selected for the current state of the network node. The method comprises comparing the weighted sum of association value of the at least one network parameter value with an weighted sum of association value of at least one other network parameter value recommended by a second application among the plurality of applications for the same at least one network parameter. If the weighted sum of association value of the at least one network parameter value is smaller than the weighted sum of association value of the at least one other network parameter value recommended by the second application, the method comprises determining to configure the at least one network parameter using the recommended at least one network parameter value. If the weighted sum of association value of the at least one network parameter value is greater than the weighted sum of association value of the at least one other network parameter value recommended by the second application, the method comprises determining not to configure the at least one network parameter using the recommended at least one network parameter value.

[0027] In some embodiments, the step of initiating the configuration of the at least one network parameter value comprises adjusting the at least one network parameter using the recommended at least one network parameter value.

[0028] In some embodiments, the step of requesting the first application for at least one alternate network configuration comprises requesting the first application to reduce a search space of the at least one network parameter value.

[0029] According to a second aspect of the present disclosure, an apparatus of a network node configured for handling network parameter recommendations from a plurality of applications is provided. The plurality of applications are executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automated platform. The network node is in communication with the automation platform. The apparatus comprises a controlling circuitry configured to cause reception of a network parameter recommendation from a first application among the plurality of applications. The network parameter recommendation comprises at least one recommended network parameter value for configuring at least one network parameter of the network node. The apparatus is configured to cause determination of whether or not to configure the at least one network parameter using the recommended at least one network parameter value. The apparatus is configured to cause initiation of configuration of the at least one recommended network parameter value when it has been determined to configure the at least one network parameter using the recommended at least one network parameter value. The apparatus is configured to cause requesting of the first application to recommend at least one alternative network parameter value when it has been determined not to configure the at least one network parameter using the recommended at least one network parameter value. A third aspect is a network node comprising the apparatus of the second aspect.

[0030] According to a fourth aspect of the present disclosure, there is provided a computer program product comprising a non-transitory computer readable medium, having thereon a computer program comprising program instructions. The computer program is loadable into a data processing unit and configured to cause execution of the method according to the first aspect when the computer program is run by the data processing unit.

[0031] In some embodiments, any of the above aspects may additionally have features identical with or corresponding to any of the various features as explained above for any of the other aspects.

[0032] An advantage of some embodiments is that alternative and / or improved approaches are provided for detecting and resolving conflicts that may arise when an application recommends new network parameter values for network parameters of the network node that conflict with network parameter values recommended by other applications for the same network parameters of the network node.

[0033] An advantage of some embodiments is that the conflicts may be resolved before configuring the recommended network parameter values in real-time.

[0034] An advantage of some embodiments is that a high volume of network parameter values recommended for the network parameters may be automatically handled without any conflicts.

[0035] An advantage of some embodiments is that a systematic approach may be provided to handle KPI metrics associated with each of the different network parameters.

[0036] An advantage of some embodiments is that configuration time / actuation time of the applications may be prioritized (that is the applications may be scheduled based on the state of the network node and conflicting behavior of the applications) for mitigating the conflicts in configuring the recommended network parameter values.

[0037] Other advantages may be readily apparent to one having skill in the art. Certain embodiments may have none, some, or all of the recited advantages.

[0038] BRIEF DESCRIPTION OF THE DRAWINGS The foregoing will be apparent from the following more particular description of the example embodiments, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the example embodiments.

[0039] Fig. 1 discloses a wireless communication system according to some examples;

[0040] Fig. 2 is a flowchart illustrating method steps according to some examples;

[0041] Fig. 3 is a flowchart illustrating method steps according to some examples;

[0042] Fig. 4 is a flowchart illustrating method steps according to some examples;

[0043] Fig. 5 is a flowchart illustrating method steps according to some examples;

[0044] Fig. 6 discloses an example lookup tables, LUTs, built for a number of possible states of a network node;

[0045] Fig. 7 is a schematic block diagram illustrating an example apparatus;

[0046] Fig. 8 discloses an example plot comprising an objective function value (targeted KPI metric) and all other Key Performance Indicator, KPI, metrics at each time step in a network node;

[0047] Fig. 9 discloses an example illustration of scheduling of applications; and

[0048] Fig. 10 discloses an example computing environment.

[0049] DETAILED DESCRIPTION

[0050] Aspects of the present disclosure will be described more fully hereinafter with reference to the accompanying drawings. The apparatus and method disclosed herein can, however, be realized in many different forms and should not be construed as being limited to the aspects set forth herein. Like numbers in the drawings refer to like elements throughout.

[0051] The terminology used herein is for the purpose of describing particular aspects of the disclosure only and is not intended to limit the invention. It should be emphasized that the term "comprises / comprising" when used in this specification is taken to specify the presence of stated features, integers, steps, or components, but does not preclude the presence or addition of one or more other features, integers, steps, components, or groups thereof. Fig. 1 discloses an example wireless communication system 100. Although the subject matter described herein may be implemented in any appropriate type of system using any suitable components, the examples disclosed herein are described in related to a wireless communication system / wireless network, such as the example wireless communication system 100 described in Fig. 1.

[0052] The wireless communication system 100 may comprise and / or interface with any type of communication, telecommunication, data, cellular, and / or radio network or other similar type of system.

[0053] In some examples, the wireless communication system 100 may be configured to operate according to specific standards or other types of predefined rules of procedures. Thus, particular embodiments of the wireless communication system 100 may implement communication standards, such as, but are not limited to, global system for mobile communications, GSM, universal mobile telecommunications system, UMTS, long term evolution, LTE, and / or other suitable 2G, 3G, 4G, or 5G standards, wireless local area network, WLAN, standards such as, IEEE 802.11 standards, and / or any other appropriate wireless communication standards, such as, worldwide interoperability for microwave access, WiMax, Bluetooth, Z-Wave and / or ZigBee standards.

[0054] In some examples, an Open Radio Access Network, O-RAN, architecture or a cloud RAN architecture may be implemented within the wireless communication system 100. For simplicity, the wireless communication system 100 with the O-RAN architecture is disclosed in Fig. 1. As disclosed in Fig. 1, the wireless communication system 100 comprises an automation platform 102 and a RAN 112.

[0055] The RAN 112 comprises one or more network nodes 114 in communication with the automation platform 102. A network node 114 may connect User Equipments, UEs, to other parts, for example, a Core Network, CN (not shown in Fig. 1) of the wireless communication system 100.

[0056] In some examples, the network node 114 referred herein may include an O-RAN network node. In some examples, the network node 114 referred herein may include, but are not limited to, access points, APs (for example, radio access points), base stations, BSs (for example, radio base stations, nodeBs, evolved NodeBs, eNBs, new radio, NR, nodes (gNBs), or the like). The BSs may be categorized based on an amount of coverage the BSs provide (or, stated different, their transmit power level) and may then also be referred to as femto BSs, pico BSs, micro BSs, macro BSs. The BS may be a relay node or a relay donor node controlling a relay.

[0057] With the O-RAN architecture, functionality of the network node 114 may be distributed into one or more units, such as, a centralized unit, CU, a distributed unit, DU, and a radio unit, RU. The CU may be a logical node for hosting Packet Data Convergence Protocol, PDCP, sublayers of the network node 114. The DU may be a logical node hosting Radio Link Control, RLC, Media Access Control, MAC, and Physical, PHY, sublayers of the network node 114. The RU may be a physical node that converts radio signals from antennas to digital signals that may be transmitted over a front haul to the DU.

[0058] Further, a plurality of network parameters may be defined for the network node 114, which enables the network node 114 to communicate with the UEs, other network nodes, or other parts of the wireless communication system 100 (for example, a Core Network, CN). Examples of the network parameters may include, but are not limited to, transmission power, antenna tilt, a cell operating parameter (for example, to turn ON / OFF a cell), a handover related parameter, Cell Individual Offset, CIO, and so on. Each network parameter may be associated with one or more Key Performance Indicator, KPI, metrics. The KPI metrics associated with the network parameter may depict results derived from configurations of that network parameter. Examples of the KPI metrics may include, but are not limited to, capacity, coverage, power consumption, UE / network node fairness, and so on.

[0059] In addition, the network node may operate in different states. The states referred herein may capture different characteristics of the network node. A state of the network node at a time may be defined based on a combination of the KPI metrics and / or the respective network parameter.

[0060] The automation platform 102 may be configured to improve performance of the wireless communication system 100 and to enhance customer / end-user experience while delivering operational savings through industrial scale automation in the RAN 112. The automation platform 102 may further enable Communication Service Providers, CSPs, to deploy applications to optimize the performance of the wireless communication system 100 by reducing complexity and operational costs. The automation platform 102 may be deployed in a cloud environment.

[0061] The automation platform 102 comprises a Service Management Orchestration, SMO 104, a plurality of applications 106a-106n, and a non-real time RAN Intelligent Controller, RIC 108. It should be understood that the automation platform 102 may comprise one or more other components such as, but are not limited to, a near-real time RIC, applications designed to execute on the near-real time RIC (referred to as xApps), a cloud orchestration module, RAN Operation, Administration, and Maintenance, 0AM, module, or the like (not shown in Fig. 1).

[0062] The SMO 104 may perform management and orchestration of components of the RAN 112, which are under its control. The SMO 104 may terminate an 01 interface, an 02 interface, and an Al interface in the non-real RIC 108. The SMO 104 may also encapsulate a R1 interface between the non-real time RIC 108 and the plurality of applications 106a-106n.

[0063] The non-real time RIC 108 may be a controller dedicated for executing, managing and orchestrating the plurality of applications 106a-106n that optimize performance of the network nodes 114. The non-real time RIC 108 may control an automation loop of one second and longer.

[0064] The plurality of applications 106a-106n referred herein may also be referred to as rApps, non- real time RIC applications, or the like. The plurality of applications 106a-106n may be software applications developed by at least one of: network vendors, CSPs, specialist developers to automate and optimize performance of the network nodes 114 in the RAN 112.

[0065] In some examples, the automation platform 102 may enable each of the plurality of applications 106a-106n to obtain subscription for one or more topics with read and / or write access. With such a subscription, an application may acquire information about the one or more network nodes 114 in the RAN and / or to provide network parameter recommendations for configuring the network parameters of the one or more network nodes 114.

[0066] In order to obtain the subscription, the application may register with the automation platform 102 by providing details such as a list of intent, network parameter values that the application can recommend, and KPI metrics targeted by the application. Consider an example scenario, wherein an application 106a is deployed in the automation platform 102 to handle energy optimization. In such a scenario, the application 106a may register with the automation platform 102 to obtain a subscription for an energy optimization topic. While registering, the application 106a may provide the following details:

[0067] - energy optimization application: an intent is to maximize energy savings in the network node 114;

[0068] - network parameter values: the network parameter values may be recommended for transmission power (an example of the network parameter) of the network node 114; and

[0069] - KPI metrics: energy efficiency of cells associated with the network node 114.

[0070] Consider another example scenario, wherein an application 106b is deployed in the automation platform 102 to handle cell coverage optimization. In such a scenario, the application 106b may register with the automation platform 102 to obtain subscription for a cell coverage topic. While registering, the application 106b may provide the following details:

[0071] - cell coverage optimization application: an intent is to maximize area of coverage;

[0072] - network parameter values: the network parameter values may be recommended for antenna tilt (an example of the network parameter) of the network node; and

[0073] - KPI metrics: a number of UEs associated with a signal strength above a pre-defined threshold (for example, good signal strength).

[0074] Thus, in accordance with the subscription, each of the plurality of applications 106a-106n may provide network parameter recommendations comprising network parameter values for configuring the network parameters of the network node 114 while targeting the one or more KPI metrics.

[0075] In the wireless communication system 100, upon receiving the network parameter values from the plurality of applications 106a-106n, the network node 114 may configure the network parameter values in accordance with a first-in -first-out, FIFO, scheme and independent of any quality measures associated with the network parameter values. This may cause conflict in configuring the network parameter values received from the plurality of applications 106a-106n, when the plurality of applications 106a-106n targets optimizing the different KPI metrics associated with the same network parameters of the network node 114. Consider an example scenario, wherein an application 106a recommends a network parameter value for an antenna tilt of the network node 114 by targeting a first KPI metric, for example, capacity. However, the network parameter value recommended for the antenna tilt may also target a second KPI metric, for example, a coverage overlap area between neighbour cells. This may affect a handover success rate KPI metric, since an application 106b has recommended a network parameter value (for example, a hysteresis threshold value) for a handover related parameter of the network node 114 without considering the change in the coverage overlap area (i.e., a new shape of cell). This may lead to a conflict between the network parameter values generated by the different applications. The conflict may further lead to performance degradation of the network node 114, which may have disastrous results.

[0076] Therefore, the network node 114 herein implements a method for handling the network parameter recommendations generated by the plurality of applications without any conflicts, thereby performance degradation of the network node 114 may be reduced.

[0077] The network node 114 receives a network parameter recommendation from at least an application, for example, a first application 106a among the plurality of applications 106a- 106n. The network parameter recommendation comprises at least one recommended network parameter value for configuring at least one network parameter of the network node 114.

[0078] Upon receiving the network parameter recommendation, the network node 114 determines whether or not to configure the at least one network parameter using the recommended at least one network parameter value.

[0079] When it has been determined to configure the at least one network parameter using the recommended at least one network parameter value, the network node 114 initiates configuration of the at least one recommended network parameter value. When it has been determined not to configure the at least one network parameter using the recommended at least one network parameter value, the network node 114 requests the first application 106a to recommend at least one alternative network parameter value. Various examples for handling the network parameter recommendations from the plurality of applications 106a-106n are explained in conjunction with figures in the later parts of the description.

[0080] Fig. 2 is a flowchart illustrating example method steps of a method 200 performed by a network node for handling network parameter recommendations from a plurality of applications. The plurality of applications referred herein are applications (also referred to as rApps) executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automation platform. The network node is in communication with the automation platform.

[0081] At step 202, the method 200 comprises receiving a network parameter recommendation from a first application among the plurality of applications. The network parameter recommendation comprises at least one recommended network parameter value for configuring at least one network parameter of the network node. The at least one network parameter of the network node may be a parameter configured for enabling the network node to communicate with User Equipments, UEs, or other network nodes, or other components of a wireless communication system. Examples of the at least one network parameter may include, but are not limited to, antenna tilt, power consumption, transmission power, cell operating parameters, handover / mobility related parameters, Cell Individual Offset, CIO, UE distribution parameters, and so on. Each network parameter may be associated with a plurality of Key Performance Indicator, KPI, metrics. The KPI metrics associated with the network parameter may depict results derived from configurations of that network parameter. Examples of the KPI metrics may include, but are not limited to, capacity, coverage, power consumption, UE / network node fairness, and so on.

[0082] At step 204, the method 200 comprises determining whether or not to configure the at least one network parameter using the recommended at least one network parameter value.

[0083] In some embodiments, the step 204 may comprise identifying the plurality of KPI metrics associated with the at least one network parameter for which the at least one network parameter value is recommended. The method may comprise determining at least one KPI metric from the plurality of KPI metrics of the at least one network parameter targeted by the recommended at least one network parameter value. In some examples, the network node may determine the at least one KPI metric targeted by the recommended at least one network parameter value based on a subscription of the first application. If the first application has a subscription to an energy optimization, the network node may determine that the at least one KPI metric targeted by the recommended at least one network parameter value is power consumption. Similarly, if the first application has a subscription to antenna tilt, the network node may determine that the at least one KPI metric targeted by the recommended at least one network parameter value is coverage.

[0084] Upon determining the targeted KPI metric, the method may comprise determining whether there exists any conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value.

[0085] When it has been determined that there is no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, the method 200 may comprise determining to configure the at least one network parameter usingthe recommended at least one network parameter value. When it has been determined that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, the method 200 may comprise determining not to configure the at least one network parameter using the recommended at least one network parameter value.

[0086] Consider an example scenario, wherein the network node receives, from the first application, a network parameter value recommended for antenna tilt (example network parameter) of the network node. In such a scenario, the network node determines that coverage is a KPI metric targeted by the recommended network parameter value, based on a subscription of the first application. The network node further determines whether there exists conflict with other KPI metrics (such as capacity, power consumption, fairness, or the like) from the recommended network parameter value. If there exists no conflict with the other KPI metrics, the network node determines to configure the antenna tilt using the recommended network parameter value. If there exists a conflict with the other KPI metrics, the network node determines not to configure the antenna tilt using the recommended network parameter value. Step 204 is described in detail in conjunction with Figs. 3 and 4.

[0087] When it has been determined to configure the at least one network parameter using the recommended at least one network parameter value, at step 206, the method 200 comprises initiating configuration of the at least one recommended network parameter value. In some embodiments, the step 206 of initiating the configuration of the at least one recommended network parameter value may comprise adjusting the at least one network parameter using the recommended network parameter value.

[0088] When it has been determined not to configure the at least one network parameter using the recommended at least one network parameter value, at step 206, the method 200 comprises requesting the first application to recommend at least one alternative network parameter value. In some embodiments, the step 206 of requesting the first application to recommend at least one alternative network parameter value may comprise requesting of the application to reduce a search space of the at least one network parameter value. Reducing the search space may refer to determining a difference between the recommended at least one network parameter value and at least one current network parameter value of the at least one network parameter and providing the at least one alternative network parameter value in accordance with the determined difference.

[0089] Consider an example scenario, wherein the first application recommends the network parameter recommendation comprising three network parameter values to increase downtilt of a cell by 2 units, increase downlink transmission power by one unit, and turn ON a small cell. The three network parameter values are targeting a coverage metric (an example of the KPI metrics) associated with the three network parameters. In such a scenario, when it has been determined that there exists no conflict with the other KPI metrics from the recommended three network parameter values, the network node may adjust the down tilt of the cell by 2 units, increase the downlink transmission power by one unit, and turn ON the small cell.

[0090] In some examples, the network node may configure one of the recommended network parameter values at a time. That is the network node may configure the recommended network parameter values sequentially, when there exists no conflict with the other KPIs of the network parameter from all the recommended network parameter values. Thus, the network node may be given time to stabilize to observe a final effect of the configuration on the KPI metrics associated with the respective network parameter.

[0091] When it has been determined that there exists a conflict with the other KPI metrics from the recommended second and third network parameter values, the network node may adjust only the down tilt of the cell by 2 units. Thereby, the network node may initiate configuration of any of the recommended network parameter values, which does not cause any conflict with the other KPI metrics.

[0092] When it has been determined that there exists conflict with the other KPI metrics from all the recommended three network parameter values, the network node may request the application to reduce a search space of the recommended three network parameter values or to recommend new three network parameter values.

[0093] Therefore, the network parameter values may be configured without any conflicts and with minimum configuration time.

[0094] Fig. 3 is a flowchart illustrating example method steps 204 performed by a network node for configuring of at least one recommended network parameter value. The network node receives a network parameter recommendation from a first application among a plurality of applications executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automation platform. The network node is in communication with the automation platform. The network parameter recommendation comprises at least one network parameter value for at least one network parameter of the network node. Examples of the at least one network parameter may include, but are not limited to, antenna tilt, power consumption, transmission power, cell operating parameters, handover / mobility related parameters, Cell Individual Offset, CIO, UE distribution parameters, and so on. Each network parameter may be associated with a plurality of Key Performance Indicator, KPI, metrics. The KPI metrics associated with the network parameter may depict results derived from configurations of that network parameter. Examples of the KPI metrics may include, but are not limited to, capacity, coverage, power consumption, UE / network node fairness, and so on.

[0095] Upon receiving the network parameter recommendation, the network node determines, at step 204, whether to configure or not the at least one network parameter using the recommended at least one network parameter value. Step 204 comprises sub steps 302-310. At step 302, the network node identifies the plurality of KPI metrics associated with the at least one network parameter for which the at least one network parameter value is recommended.

[0096] In some examples, the network node may pre-define a KPI metric list for each of a plurality of network parameters of the network node based on previous results of configurations of that network parameter. The KPI metric list may indicate the KPI metrics associated with each network parameter. It should be understood that the KPI metric list may also be changed, based on configuration results of the different network parameters of the network node. The network node may use the KPI metric list to identify the KPI metrics for the at least one network parameter for which the at least one network parameter value is recommended.

[0097] At step 304, the network node determines at least one KPI metric from the plurality of KPI metrics of the at least one network parameter targeted by the recommended at least one network parameter value. In some examples, each application may recommend the at least one network parameter value to optimize targeted one or more KPI metrics associated with the respective network parameter of the network node. Each application may target the one or more KPI metrics based on its subscription. For example, the first application may target coverage, a second application may target power consumption, a third application may target all the KPI metrics associated with the at least one network parameter of the network node, or the like. Therefore, the network node may identify the targeted at least one KPI metric based on the subscription of the first application recommending the at least one network parameter value.

[0098] At step 306, the network node determines whether there exists any conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value. It should be understood that determining whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value may refer to determining whether the recommended at least one network parameter value changes / impacts / effects the other KPI metrics of the at least one network parameter.

[0099] For determining whether there exists any conflict with the other KPI metrics, the network node may select, from lookup tables, LUTs, built for a number of possible states of the network node, a LUT for a current state of the network node for which the at least one network parameter value is recommended. The state of the network node may represent characteristics of the network node.

[0100] In some examples, the state of the network node at a specific time may be determined by a combination of the network parameters of the network node and / or the respective KPI metrics associated with the network parameters at that time. An example of the state may be (0.8, 0.5, 0.3), where 0.8, 0.5 and 0.3 are loads / utilization related KPI metrics associated with cells 1, 2, and 3 in the network at a particular time.

[0101] The LUT may comprise entries for each cell, and each network parameter associated with one or more KPI metrics for each cell. For example, the LUT built forthe state of the network node with respect to a cell of the network node may comprise the network parameters, the KPI metrics associated with each network parameter, and an association value for each combination of a network parameter and a respective KPI metric. The association value (also be referred to as impact value, impact score, or the like) may indicate an association of the KPI metric with the respective network parameter. Consider an example scenario, wherein network parameters of the network node, for example, antenna tilt, and transmission power are associated with KPI metrics for example, capacity and power consumption. In such a scenario, the LUT may comprise an association value of 0.7 for a combination of the antenna tilt and the coverage, 0.1 for a combination of the antenna tilt and the power consumption, 0.6 for a combination of the transmission power and the power consumption, and 0.9 for a combination of the transmission power and the power consumption.

[0102] In some examples, the LUT may also comprise priorities / weights for the KPI metrics associated with each of the network node. In some examples, the priorities for the KPI metrics may be pre-defined based on Service Level Agreements, SLAs, agreed with network operators. In some examples, the priorities for the KPI metrics may be defined based on some function of the targeted KPI metric and the current KPI metrics. Example of such function may be a difference between the targeted and current KPI metrics or a difference between a minimum and current KPI metrics. In some examples, the priorities for the given LUT may be same for all the applications. In some examples, prior to selecting the LUT for the current state of the network node, the network node may identify the number of possible states of the network node. Each possible state may represent different characteristics of the network node in terms of the plurality of network parameters and the KPI metrics associated with each network parameter. For example, consider that the network parameter is a UE distribution. In such a scenario, the possible states of the network node may include a low load state, a medium load state, and a high load state.

[0103] In some examples, the network node may identify the number of possible states when there is any change in the network parameters defined for the network node (that is defining of a new network parameter for the network node) or when at least one new KPI metric has been defined for any of the network parameters.

[0104] For identifying the number of possible states of the network node, the network node may determine a maximum possible number of states of the network node, as the states of the network node may be infinite. The network node may determine the maximum possible number of states of the network node by evaluating each of the plurality of network parameters of the network node and the respective KPI metrics using a first learning model. The first learning model is described later in Fig. 3. Upon determining the maximum possible number of states of the network node, the network node may built the LUT for each state of the maximum possible number of states. The network node may further group the states of the network node when the LUTs associated with that states differ from each other by a predefined value. For example, the pre-defined value may be represented as epsilon, e. The grouped states may be identified as the number of possible states of the network node.

[0105] Upon identifying the number of possible states of the network node, the network node may build the LUT for each state of the network node.

[0106] In some examples, the LUT has to be built for each state of the network node, since the association of each network parameter with the KPI metrics (for example, tilt of cell 1 on capacity of cell 2) may change depending upon a coupling factor among cells (for example, interference), which in turn depends on the state of the network node. For example, if the state of the network node represents loads of the cell, then high load of the cell may lead to high interference to neighbour cells and vice versa. Thus, effectively for each state of the network node (dependent on the loads of the cell), the network node may built the LUT of association values / impact values with dimensions as number of network parameters times number of KPI metrics.

[0107] For building the LUT for each state of the network node, the network node may obtain historical data collected for a pre-determined time interval. In some examples, the predetermined time interval may include days, weeks or months. It should be understood that only a shorter period of time interval may be considered for pre-defining the time interval. In some examples, the historical data may comprise a plurality of network parameter values recommended by the plurality of applications, a state of the network node associated with at least one of the recommended plurality of network parameter values, at least one network parameter configured using at least one of the plurality of network parameter values, and the KPI metrics resultant from the configuration of that at least one network parameter.

[0108] Upon obtaining the historical data, the network node may determine the association value for each combination of one of the plurality of network parameters and one of the respective KPI metrics. Thereby, the LUT may be built for the state of the network node with the association values.

[0109] In some examples, the network node may determine the association value for each combination of one of the plurality of network parameters and one of the respective KPI metrics based on evaluation of the historical data. The historical data comprising the network parameter values may be instantaneous, as it is event driven. That is a record may be generated / logged when the network parameter values have been changed. However, the historical data comprising the KPI metrics may be often aggregated in time intervals. Therefore, high resolution user level and cell level traces and Minimization of Drive Test, MDT, reports may be used to record the KPI metrics. Such traces and / or MDT reports may be combined together with the changes in the network parameter values to determine the association or impact of network parameter values on the KPI metrics.

[0110] In some examples, the network node may determine the association value for each combination of one of the plurality of network parameters and one of the respective KPI metrics, by evaluating the historical data using a second learning model. The second learning model is described later in Fig. 3. In some examples, the second learning model may evaluate the historical data by identifying at least one feature as importance from the obtained historical data. The second learning model may be a decision tree fitted on the historical data by considering the network parameters as input features and the KPI metrics as labels. Upon fitting the decision tree, the at least one feature as importance (also be referred to as feature importance) may be identified by determining how much each input feature (i.e., network parameter) is contributing to reducing uncertainty in the target KPI / variable. The feature importance may be used as association values in the LUT. Alternatively, the second learning model may calculate SHapely Additive exPlanation, SHAP, values / scores, which may use a game theoretic approach to measure each input feature's contribution to a final output of the second learning model. The SHAP values / scores may be used as the association values in the LUT.

[0111] In some examples, the second learning model may evaluate the historical data based on interpolation between the states of the network node in which the states have been arranged in an order. In some examples, the second learning model may evaluate the historical data based on similarity measures between the states of the network node in which the states have not been arranged in the order. In some examples, the second learning model may evaluate the historical data based on an input received from a radio domain expert. The input may include the association values determined by the radio domain expert.

[0112] Thereby, the LUTs may be built for the number of possible states of the network node based on the evaluation of the historical data using the second learning model. The network node may store the LUTs built for the number of possible states of the network node in a database or in a memory. Further, the network node may update the LUTs. In some examples, the network node may update the LUTs after configuring the recommended at least one network parameter value corresponding to the respective states of the network node. In some examples, the network node may update the LUTs based on update information. The network node may receive the update information related to at least one application from the automation platform. The update information indicates at least one of: on-boarding of a new application, and a subscription of the at least one application targeting new KPI metrics. The network node may update the LUTs in accordance with the received update information.

[0113] The network node may select, from the LUTs, the LUT for the current state of the network node. For selecting the LUT for the current state of the network node, the network node may detect, using a distance metric, whether any of the identified number of states of the network node matches with the current state of the network node for which the at least one network parameter value is recommended. In some examples, the network node may compute a state vector corresponding to the current state of the network node. Similarly, the network node may derive the stored LUTs, identify the number of possible states associated with the LUTs, and compute state vectors based on the identified number of possible states. The network node may compare the state vector corresponding to the current state of the network node with the state vectors corresponding to the identified number of possible states using a distance metric (for example, Euclidian space). Based upon the comparison, the network node may identify the state vector among the state vectors corresponding to the identified number of possible states that is nearest to the state vector corresponding to the current state in Euclidian space. The network node may determine the state associated with the identified state vector as the state of the identified number of states of the network node matching with the current state of the network node.

[0114] When it has been determined that one of the identified number of states of the network node matches with the current state of the network node, the network node may select the LUT corresponding to that state as the LUT for the current state of the network node.

[0115] Upon selecting the LUT, the network node may determine using the selected LUT, whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value.

[0116] The network node may use the selected LUT and determine a weighted sum of association value of the recommended at least one network parameter value with respect to each of the other KPI metrics different from the targeted KPI metric of the at least one network parameter.

[0117] In some examples, for determining the weighted sum of association value of the recommended at least one network parameter value, the network node may identify from the selected LUT, a priority and an association value associated with each of the other KPI metrics with respect to the at least one network parameter. The network node may determine the weighted results for the other KPI metrics based on the priority and the association value of each of the other KPI metrics. Determining a weighted result for other KPI metric may refer to multiplication of the association value associated with that KPI metric with a priority of that KPI metric. The network node may sum up the weighted results to determine the weighted sum of association value of the recommended at least one parameter value. Determining the weighted sum of association value of the recommended at least one network parameter value is described in detail with examples in conjunction with Fig. 5.

[0118] The network node may compare the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics with a threshold factor. The threshold factor may be defined based on the current state of the network. In some examples, the network node may tune the threshold factor dynamically. For tuning the threshold factor, the network node may evaluate the historical data using a third learning model to determine the at least one network parameter configured with a minimum configuration time and a minimum impact on other KPI metrics different from the targeted KPI metric of the at least one network parameter. The network node may tune the threshold factor based on the evaluation.

[0119] When the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics is smaller than the threshold factor, the network node may determine that there exists no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value.

[0120] When the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics is greater than the threshold factor, the network node may determine that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value.

[0121] When it has been determined that there is no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, at step 308, the network node may determine to configure the at least one network parameter using the recommended at least one network parameter value.

[0122] When it has been determined that there is no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, at step 310, the network node may determine not to configure the at least one network parameter using the recommended at least one network parameter value.

[0123] It should be understood that each of the first learning model, the second learning model, and the third learning model described in Fig. 3 may include but are not limited to, an Artificial Intelligence, Al, model, a neural network model, a machine learning model, a multi-class support vector machine, SVM, model, a recurrent neural network, RNN, model, a restricted Boltzmann machine, RBM, model, a deep belief network, DBN, model, a generative adversarial network, GAN, model, a regression based neural network model, a deep reinforcement model, a deep Q-network model, and so on. Each learning model may include a plurality of nodes arranged in layers. Examples of the layers may include, but are not limited to, a convolutional layer, a concatenated layer, a dropout layer, a fully connected layer, a SoftMax layer, and so on. Each layer has weights and performs a layer operation through calculation of a previous layer and an operation of a plurality of weights / coefficients.

[0124] In some examples, each learning model may be trained by applying one or more learning methods on the historical data collected for the pre-defined time interval. In some examples, the learning methods may include one or more of: a supervised learning method, an unsupervised learning method, a semi-supervised learning method, a reinforcement learning method, and a regression-based learning based method. The trained learning model may include a known and fixed number of layers, and sequence for processing the layers and parameters related to each layer. The parameters may include one or more of: activation functions, biases, input weights, output weights, and so on, related to the layers. A function associated with the learning method may be performed through a memory, and a controlling circuitry. The controlling circuitry may include one or more processors such as, a central processing unit, CPU, an application processor, AP, a graphics processing unit, GPU, a visual processing unit, VPU, an Al dedicated processor (like neural processing unit, NPU), and so on.

[0125] Fig. 4 is a flowchart illustrating example method steps 204 performed by a network node for configuring of at least one recommended network parameter value. The network node receives a network parameter recommendation from a first application among a plurality of applications executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automation platform. The network parameter recommendation comprises at least one network parameter value for at least one network parameter of the network node. Upon receiving the network parameter recommendation, the network node determines, at step 204, whether or not to configure the at least one network parameter using the recommended at least one network parameter value. Step 204 comprises sub steps 402-406.

[0126] At step 402, the network node comprises determining a weighted sum of association value of the recommended at least one network parameter value using a LUT selected for a current state of the network node. Determining the weighted sum of association value of the recommended at least one network parameter value is described in conjunction with Figs. 3 and 5, therefore repeated description is omitted herein.

[0127] At step 404, the network node comprises comparing the weighted sum of association value of the at least one network parameter value with an weighted sum of association value of at least one other network parameter value recommended by a second application among the plurality of applications for the same at least one network parameter.

[0128] Based upon the comparison, at step 406, the network node determines whether or not to configure the at least one network parameter using the recommended at least one network parameter value.

[0129] The network node may determine to configure the at least one network parameter using the recommended at least one network parameter value, if the weighted sum of association value of the at least one network parameter value is smaller than the weighted sum of association value of the at least one other network parameter value recommended by the second application.

[0130] The network node may determine not to configure the at least one network parameter using the recommended at least one network parameter value, if the weighted sum of association value of the at least one network parameter value is greater than the weighted sum of association value of the at least one other network parameter value recommended by the second application.

[0131] Fig. 5 is a flowchart illustrating example method steps performed by a network node for handling network parameter recommendations. The network node may perform steps 501- 504 during a training phase in order to build lookup tables, LUTs, for a number of possible states of the network node. The network node may perform steps 505-511 during an inference phase in order to handle network parameter recommendations received from applications using one of the built LUTs in the training phase. The applications referred herein are executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automation platform, which is in communication with the network node.

[0132] At step 501, the network node obtains historical data collected for a pre-determined time interval. In some examples, the historical data may comprise a plurality of network parameter values recommended by a plurality of applications, a state of the network node associated with at least one of the recommended plurality of network parameter values, at least one network parameter configured using at least one of a plurality of network parameter values, and Key Performance Indicator, KPI, metrics resultant from the configuration of said at least one network parameter. Examples of the at least one network parameter may include, but are not limited to, antenna tilt, power consumption, transmission power, cell operating parameters, handover / mobility related parameters, Cell Individual Offset, CIO, UE distribution parameters, and so on. The KPI metrics associated with the network parameter may depict results derived from configurations of that network parameter. Examples of the KPI metrics may include, but are not limited to, capacity, coverage, power consumption, UE / network node fairness, and so on.

[0133] At step 502, the network node identifies a maximum number of possible states of the network node. In some examples, the network node may fetch the plurality of network parameters of the network node and the respective KPI metrics from the historical data. The network node may evaluate each of the plurality of network parameters and the respective KPI metrics using a first learning model. Based upon the evaluation, the network node may identify the maximum number of possible states of the network node.

[0134] At step 503, the network node builds the LUTs for the identified maximum number of states of the network node. Building the LUT for the state of the network node may refer to finding and filling an association value for each combination of one of the network parameters and one of the respective KPI metrics. In some examples, the network node may find the association value based on evaluation of the historical data using a second learning model. The second learning model may perform SHAP value analysis to fit a decision tree on the network parameters of the historical data as input features and the respective KPIs as labels. From the decision tree, the second learning model may identify a feature of importance from the historical data and accordingly the network node may find the association value for each combination of the network parameter and the respective KPI.

[0135] At step 504, the network node groups / merges the states of the network node when the LUTs corresponding to such states differ from each other by a pre-defined value, for example, 'e'. For instance, the network node may group the states of the network node when the LUTs corresponding to such states differ from each other by small 'e'.

[0136] After merging the states of the network node, at step 504, the network node pre-processes the LUTs. The pre-processed LUTs for the identified number of possible states of the network node are disclosed in Fig. 6.

[0137] In some examples, consider a scenario, wherein all the network parameters in the historical data have not been configured in all the states of the network node. In such a scenario, the network node may find the association value for each combination of one of the network parameters and one of the respective KPI metrics using a function of tuple (network parameter, KPI) of the LUTs from a closest states of the network node. For example, if states of the network node are ordered as low load, medium load, and high load, the network node may find the association value for each combination of one of the network parameters and one of the respective KPI metrics based on evaluation of interpolation of the states using the second learning model. For another example, if states of the network node are not ordered, the network node may find the association value for each combination of one of the network parameters and one of the respective KPI metrics based on evaluation of similarity measures associated with the states using the second learning model. For another example, the network node may find the association value for each combination of one of the network parameters and one of the respective KPI metrics based on an input received from a radio domain expert. The input may indicate possible association values determined by the radio domain expert.

[0138] At step 505, the network node receives a network parameter recommendation from a first application among the plurality of applications. The network parameter recommendation comprises at least one network parameter value recommended for changing / configuring the at least one network parameter of the network node by targeting at least one KPI metric associated with the at least one network parameter. Upon receiving the network parameter recommendation, at step 506, the network node selects, from the built LUTs, a LUT for a current state of the network node for which the at least one network parameter value is recommended. The network node may determine the closest state for the current state of the network node using a distance metric. In some examples, the distance metric may be a Euclidian distance / space. The network node may further select the LUT built for the determined closest state as the LUT for the current state of the network node.

[0139] After selecting the LUT for the current state of the network node, at step 507, the network node determines, using the selected LUT, a weighted sum of association value (impact score / value) of the recommended at least one network parameter value. The network node may determine the weighted sum of association value of the recommended at least one network parameter value by finding the weighted sum of association values of the recommended at least one network parameter value for all the other KPI metrics that are not targeted by the recommended at least one network parameter value. Finding the weighted sum of association values of the recommended at least one network parameter value for all the other KPI metrics may refer to finding the weighted sum in a specific row of the LUT for all the other KPI metrics except for the targeted at least one KPI metric.

[0140] Thereafter, at step 507, the network node compares the weighted sum of association value of the recommended at least one network parameter value with a threshold factor (y). If the weighted sum of association value of the recommended at least one network parameter value is smaller than the threshold factor, the network node determines that there exists no conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value. Herein, determining there exists no conflict with the other KPI metrics may refer to determining that the recommended at least one network parameter value does not impact the other KPI metrics of the at least one network parameter. When there exists no conflict with the other KPI metrics, at step 508, the network node initiates configuration of the recommended at least one network parameter value. Initiating the configuration may refer to adjusting / changing the at least one network parameter of the network node using the recommended at least one network parameter value. If the weighted sum of association value of the recommended at least one network parameter value is greater than the threshold factor, the network node determines that there exists a conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value. Herein, determining there exists the conflict with the other KPI metrics may refer to determining that the recommended at least one network parameter value impacts the other KPI metrics of the at least one network parameter. When there exists the conflict with the other KPI metrics, at step 509, the network node requests the first application to recommend at least one alternative network parameter value. In some examples, requesting the first application to recommend the at least one alternative network parameter value may refer to requesting the first application to either reduce search space of the network parameter or suggest a new network parameter value with a smaller difference 'A'. The smaller difference 'A' may be a difference between a current network parameter value of the at least one network parameter and the recommended at least one network parameter value.

[0141] In some examples, in case if two applications recommend different network parameter values (A-L and A2) to the same network parameter by targeting the same KPI metric of the network parameter, the network node determines the weighted sum of association value of each of the different network parameter values using the LUT selected for the current state of the network. If the determined weighted sum of association value of each of the different network parameter values using the LUT is smaller than the threshold factor, the network node compares the different network parameter values between each other and determines to configure a smaller network parameter value among the different network parameter values. Otherwise, the network node may request the applications to recommend the at least one alternative network parameter value.

[0142] Consider an example scenario, wherein a cell of the network node has started to monitor a congestion that triggers a first application, which is subscribed for optimizing a first KPI metric, for example, capacity. In such a scenario, the first application receives information about the current state of the network node and recommends three network parameter values for three network parameters of the network node to optimize the capacity of the cell. For instance, the three network parameters may include antenna tilt, transmission power, and cell operating parameter (ON / OFF small cell). Further, the three network parameter values may indicate changes to stabilize the capacity of the cell by increasing down tilt of the cell by 2 units, increasing downlink transmission power by 1 unit, and turn ON the small cell.

[0143] Upon receiving the recommended three network parameter values, the network node selects the LUT for the current state of the network node. An example of the selected LUT is depicted in the below table:

[0144] Table 1: LUT

[0145] The LUT as depicted in table 1 comprises the network parameters of the network node, the KPI metrics (capacity, fairness, power consumption) for each of the network parameters, priorities of the KPI metrics (llzA2, A3) and an association value for each combination of one of the network parameters and one of the respective KPI metrics. For example, as depicted in the LUT, '0.7' may be an association value for a combination of the antenna tilt and the capacity.

[0146] Upon selecting the LUT, the network node uses the selected LUT to compute a weighted sum of association value of each of the three network parameter values with respect to fairness and power consumption, as all the three network parameter values are targeting the capacity.

[0147] For example, the network node may compute the weighted sum of association value of each of the three parameter values with respect to fairness (having the priority 'A2'= 0.5) and power consumption (having the priority 'A3' = 0.5) as follows:

[0148] Network parameter value with respect to the antenna tilt =0.8 * 12+ 0.1 * 13= 0.8 * 0.5 + 0.1 * 0.5 = 0.9

[0149] Network parameter value with respect to the transmission power = 0.4 * A2+ 0.9 * A3= 0.4 * 0.5 + 0.9 * 0.5 = 1.3 Network parameter value with respect to the transmission power =0.5 * 12+ 0.6 * 13= 0.5 * 0.5 + 0.6 * 0.5 = 1.1

[0150] The network node compares the weighted sum of association value of each of the three parameter values with the threshold factor. In an example herein, consider that the threshold factor is '1'. In accordance with the comparison, the network node determines that the network parameter value recommended for the antenna title does not impact / cause any conflict with the other KPI metrics such as the fairness and the power consumption, since the weighted sum of association value of the network parameter value recommended for the antenna title is smaller than the threshold factor. Therefore, the network node initiates configuration of the network parameter value recommended for the antenna tilt. In an example herein, initiating the configuration of the recommended network parameter value refers to increasing the down-tilt of the cell by 2 units. An example LUT updated after configuring / actuating the recommended network parameter value is depicted in the below table: parameter value

[0151] The network node may further calculate the weighted sum of association value of second and third network parameter values recommended for the transmission power and the ON / OFF small cell (similar to as described above) using the updated LUT as depicted in table 2. For example herein, consider that the weighted sum of association value of the network parameter value recommended for the transmission power is smaller than the threshold factor. Therefore, the network node may initiate configuration of the network parameter value recommended for the transmission power. After configuration of the network parameter value recommended for the transmission power, the network node may determine that the state of the network node has not changed (for example, the network node in the same load state). In such a case, the network node may initiate configuration of the third network parameter recommended for the ON / OFF of the small cell.

[0152] In case, if the weighted sum of association value of recommended second and third network parameter values is greater than the threshold factor, the network node requests the first application to either suggest / recommend alternative network parameter values or suggest small changes in the recommended second and third network parameter values.

[0153] Optionally, as disclosed in Fig. 5, the network node may perform steps 510 and 511. At step 510, the network node tunes the threshold factor. The threshold factor may be dependent on the state of the network node. The network node may tune the threshold factor based on past conflict avoidance scenarios. For example, a larger value of the threshold factor may lead to quick configuration / actuation of the recommended network parameter value, thereby, agility or response time of the applications may be improved and optimized. However, the larger value of the threshold factor may have a higher chance of creating conflict in configuring the recommended network parameter values, which may result in performance degradation of the network node. Hence, there may be a trade-off and accordingly, the value of the threshold factor may be tuned such that the recommended network parameter values may be configured with minimum configuration / actuation time and minimum conflict.

[0154] At step 511, the network node updates the LUT(s). The network node may update the LUT after each configuration of the network parameter value and / or after reception of update information. The update information may indicate at least one of: on-boarding of new applications in the automation platform and subscription of the application(s) towards new KPI metric(s) that were not present in the LUT. In some examples, the network node may update / pre-populate the LUT based on some function of impact of the network parameter values on the KPI metric that is closest to the new KPI metric.

[0155] Fig. 7 is an example schematic diagram showing an apparatus 114. The apparatus 114 may e.g. be comprised in a network node. The apparatus 114 is capable of handling network parameter recommendations from a plurality of applications and may be configured to cause performance of the method 200 for handling the network recommendation from the plurality of applications. The plurality of applications referred herein are applications executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC, of an automation platform. The apparatus 114 may be in communication with the automation platform.

[0156] According to at least some examples of the present invention, the apparatus 114 in Fig. 7 comprises one or more modules. These modules may e.g. be a memory 702, a processor 704, a controlling circuitry 706, a transceiver 708, a determination module 710, and a configuration module 712. The controlling circuitry 706, may in some embodiments be adapted to control the above mentioned modules.

[0157] The memory 702, the processor 704, the transceiver 708, the determination module 710 and the configuration module 712, as well as the controlling circuitry 706, may be operatively connected to each other.

[0158] The controlling circuitry 706 may be adapted to control the steps as executed by the network node. For example, the controlling circuitry 706 may be adapted to handle the network parameter recommendations from the plurality of applications (as described above in conjunction with the method 200 and Fig. 2).

[0159] The transceiver 708 may be adapted to receive, from at least a first application among the plurality of applications, a network parameter recommendation. The network parameter recommendation comprises at least one network parameter value recommended for at least one network parameter of the network node. Each network parameter of the network node may be associated with a plurality of Key Performance Indictor, KPI, metrics.

[0160] The determination module 710 may be adapted to determine whether or not to configure the at least one network parameter using the recommended at least one network parameter value. The network node may determine the at least one KPI metric of the at least one network parameter targeted by the recommended at least one network parameter value. The network node may select, from lookup tables, LUTs, built for a number of possible states of the network node, a LUT for a current state of the network node. The network node may determine, using the selected LUT, a weighted sum of association value of the recommended at least one network parameter value with respect to each of other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter. The network node may determine to configure the recommended at least one network parameter using the recommended at least one network parameter value, when the weighted sum of association value of the recommended at least one network parameter value with respect to each of other KPI metrics is smaller than a threshold factor. The network node may determine not to configure the recommended at least one network parameter using the recommended at least one network parameter value, when the weighted sum of association value of the recommended at least one network parameter value with respectto each of other KPI metrics is greater than a threshold factor.

[0161] The determination module 710 may also be adapted to identify the number of possible states of the network node using a first learning model 716 and build the LUTs for the number of possible states of the network node using a second learning model 718. The determination module 710 may also be adapted to tune the threshold factor using a third learning model 720. The first, second, and third learning models 716-720 are already described in conjunction with Fig. 3, thereof repeated description is omitted herein.

[0162] The configuration module 712 may be adapted to initiate configuration of the recommended at least one network parameter value, when the determination module 710 determines to configure the at least one network parameter using the recommended at least one network parameter value.

[0163] The configuration module 712 may also be adapted to request the first application to recommend at least one alternative network parameter value, when the determination module 710 determines not to configure the at least one network parameter using the recommended at least one network parameter value.

[0164] The processor 704 may be adapted to manage configurations of the recommended at least one network parameter value.

[0165] Further, the memory 702 is adapted to store the number of possible states of the network node, the network parameters, the KPI metrics, the LUTs, or the like.

[0166] An example illustration of handling network parameter recommendations from a plurality of applications are explained in conjunction with Figs. 8 and 9. Consider an example scenario, wherein five applications (rApps) have been deployed in an automation platform for monitoring functionalities / operations of a network node and accordingly generating network parameter recommendations for network parameters of the network node. Examples of the network parameters may include, antenna tilt, transmission power, cell operating parameters, Cell Individual Offset, CIO, or the like. Each network parameter may be associated with Key Performance Indicator, KPI, metrics, such as, but are not limited to, coverage, capacity, User Equipment, UE, fairness, network node fairness, power consumption, or the like. The coverage may be defined as minimum of (percentage of total UEs having Reference Signal Received Power, RSRP > threshold (-lOOdBm), percentage of total UEs having Signal-to-lnterference-plus-Noise Ratio, SINR > threshold (0 dB)). The capacity may be defined as a sum of achievable throughputs of all the UEs accessing the network node. The UE fairness may be defined as a fairness index calculated for UE throughputs, wherein the index may take a maximum value (1) when all the UEs have the same achievable throughput. The network node fairness may be defined as fairness index calculated for network node cell loads utilization, wherein the index may take a maximum value (1) when all the network nodes have the same cell load utilization. The power consumption may be calculated as a sum of downlink transmission power of all the network nodes divided by a sum throughput of the network node.

[0167] Examples of the five applications deployed in the automation platform may include a coverage maximization rApp, a capacity maximization rApp, a UE fairness rApp, an energy saving rApp, and a gold rApp.

[0168] The coverage maximization rApp may generate a network parameter recommendation comprising a network parameter value for changing / configuring the antenna tilt of the network node, while optimizing the coverage. The coverage maximization rApp may check current coverage associated with the antenna tilt and only recommend the network parameter value if the current coverage is less than 95%. If a new network parameter value is required, then the coverage maximization rApp may generate the new network parameter value with a small probability of epsilon (10%). Otherwise, based on historical KPI metrics and configurations of the antenna tilt, the coverage maximization rApp may recommend the network parameter value that has maximum associated coverage. The capacity maximization rApp may generate a network parameter recommendation comprising a network parameter value for changing / configuring the antenna tilt of the network node, while optimizing the capacity using epsilon greedy rule. The capacity maximization rApp may check current capacity associated with the antenna tilt and only recommend the network parameter value if the current normalized capacity is less than 0.95. If a new network parameter value is required, then the capacity maximization rApp may generate the new network parameter value with a high probability of epsilon (50%). Otherwise, based on historical KPI metrics and configurations of the antenna tilt, the capacity maximization rApp may recommend the network parameter value that has maximum associated capacity.

[0169] The UE fairness rApp may generate a network parameter recommendation comprising a network parameter value for changing / configuring antenna azimuth and CIO of the network node, while optimizing UE throughputs using epsilon greedy rule. The UE fairness rApp may check current UE fairness associated with the antenna tilt and the CIO of the network node and only recommend the network parameter value if the current UE fairness is less than 0.95. If a new network parameter value is required, then the UE fairness rApp may generate the new network parameter value with a high probability of epsilon (100%). Otherwise, based on historical KPI metrics and configurations of the antenna azimuth and the CIO, the UE fairness rApp may recommend the network parameter value that has maximum associated UE fairness.

[0170] The energy saving rApp may generate a network parameter recommendation comprising a network parameter value for changing / configuring antenna transmission power of the network node, while optimizing the energy efficiency (sum of UE throughputs / sum of network node power consumption) using epsilon greedy rule. The energy saving rApp may check current power consumption associated with transmissions and only recommend the network parameter value if the current power consumption is greater than 'O'. If a new network parameter value is required, then the energy saving rApp may generate the new network parameter value with a high probability of epsilon (75%). Otherwise, based on historical KPI metrics and configurations of the transmission power, the energy saving rApp may recommend the network parameter value that has maximum energy efficiency. The gold rApp may generate network parameter recommendations comprising network parameter values for changing / configuring all the network parameters of the network node, while optimizing the multi-objective KPI metrics (coverage + capacity+ UE fairness + network node fairness + power consumption) using epsilon greedy rule. The gold rApp may check current all the KPI metrics and only recommend the network parameter value if any of the coverage, the capacity, the UE fairness, and the network node fairness is less than 0.95 or the power consumption is greater than 'O'. If a new network parameter value is required, then the gold rApp may generate the new network parameter value with a probability of epsilon. The value of epsilon may start with 1.0 and decrement by 0.1 on every cell. Otherwise, based on historical KPI metrics and configurations of the network parameters, the gold rApp may recommend the network parameter value that has maximum associated KPI metrics.

[0171] The network node may receive, from all the five rApps, the recommended network parameter values for configuring the network parameters of the network node. In some examples, in such a scenario, the network node may initiates configuration of the recommended network parameter values based on a first-in-first-out, FIFO, scheme, thereby, all the recommended network parameter values may be configured / scheduled all the time. In contrast to such configurations, the network node implements a method to determine, using a LUT correspond to a current state of the network node, whether there exists any conflict with other KPI metrics different targeted KPI metric of the network parameter from each of the recommended network parameter values. The network node may only initiate configuration of any of the recommended network parameter values, of there exists no conflict with the other KPI metrics of the network parameters from that recommended network parameter value. Thus, the recommended network parameter values from the different applications may be configured / scheduled without any conflicts.

[0172] An example plot comprising an objective function value (targeted KPI metric) and all other KPI metrics at each time step in the network node with respect to FIFO and LUT schedulers of the network node is disclosed in Fig. 8. The LUT scheduler may be equivalent to a controlling circuitry as disclosed in Fig. 7. The FIFO scheduler may configure / schedule the recommended network parameter values based on the FIFO scheme. The LUT scheduler may configure / schedule the recommended network parameter values by determining, using the LUT, whether there exists any conflict with other KPI metrics different targeted KPI metric of the network parameter from each of the recommended network parameter values. From Fig. 8, it is evident that configuring of the recommended network parameter values using the FIFO scheduler may lead to very unstable network operation, since it actuates / configures the recommended network parameter values irrespective of past results. On the contrary, configuring of the recommended network parameter values using the LUT scheduler may prioritize configuring of the recommended network parameter values from the different rApps, which may improve all the KPI metrics. In an example herein, as disclosed in Fig. 9, the gold rApp may be scheduled much more often by the LUT scheduler as compared to the FIFO scheduler.

[0173] Any appropriate steps, methods, features, functions, or benefits disclosed herein may be performed through one or more functional units or modules of one or more virtual apparatuses. Each virtual apparatus may comprise a number of these functional units. These functional units may be implemented via processing circuitry, which may include one or more microprocessor or microcontrollers, as well as other digital hardware, which may include digital signal processors, DSPs, special-purpose digital logic, and the like. The processing circuitry may be configured to execute program code stored in memory, which may include one or several types of memory such as read-only memory (ROM), random-access memory, RAM, cache memory, flash memory devices, optical storage devices, etc. Program code stored in memory includes program instructions for executing one or more telecommunications and / or data communications protocols as well as instructions for carrying out one or more of the techniques described herein. In some implementations, the processing circuitry may be used to cause the respective functional unit to perform corresponding functions according one or more embodiments of the present disclosure.

[0174] The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and / or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the scope of the disclosure.

[0175] Fig. 10 illustrates an example computing environment 1000 implementing a method, and a network node, as described in Figs. 2 and 7. As depicted in Fig. 10, the computing environment 1000 comprises at least one data processing module 1006 that is equipped with a control module 1002 and an Arithmetic Logic Unit, ALU, 1004, a plurality of networking devices 1008 and a plurality Input output, I / O devices 1010, a memory 1012, a storage 1014. The data processing module 1006 may be responsible for implementing the method described in Fig. 2. For example, the data processing module 1006 may in some embodiments be equivalent to the controlling circuitry of the network node described above in conjunction with Fig. 7. The data processing module 1006 is capable of executing software instructions stored in memory 1012. The data processing module 1006 receives commands from the control module 1002 in order to perform its processing. Further, any logical and arithmetic operations involved in the execution of the instructions are computed with the help of the ALU 1004.

[0176] The computer program is loadable into the data processing module 1006, which may, for example, be comprised in an electronic apparatus (such as the network node). When loaded into the data processing module 1006, the computer program may be stored in the memory 1012 associated with or comprised in the data processing module 1006. According to some embodiments, the computer program may, when loaded into and run by the data processing module 1006, cause execution of method steps according to, for example, any of the method illustrated in Fig. 2, or otherwise described herein.

[0177] The overall computing environment 1000 may be composed of multiple homogeneous and / or heterogeneous cores, multiple CPUs of different kinds, special media and other accelerators. Further, the plurality of data processing modules 1006 may be located on a single chip or over multiple chips.

[0178] The algorithm comprising of instructions and codes required for the implementation are stored in either the memory 1012 or the storage 1014 or both. At the time of execution, the instructions may be fetched from the corresponding memory 1012 and / or storage 1014, and executed by the data processing module 1006. In case of any hardware implementations various networking devices 1008 or external I / O devices 1010 may be connected to the computing environment to support the implementation through the networking devices 1008 and the I / O devices 1010.

[0179] The embodiments disclosed herein can be implemented through at least one software program running on at least one hardware device and performing network management functions to control the elements. The elements shown in Fig. 10 include blocks which can be at least one of a hardware device, or a combination of hardware device and software module.

Claims

CLAIMS1. A method (200) for handling network parameter recommendations from a plurality of applications (106a-106n), said plurality of applications (106a-106n) being executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC (108), of an automation platform (102), the method (200) being performed by a network node (114) in communication with the automation platform (102), the method (200) comprising:- receiving (202) a network parameter recommendation from a first application (106a) among the plurality of applications (106a-106n), said network parameter recommendation comprising at least one recommended network parameter value for configuring at least one network parameter of the network node (114);- determining (204) whether or not to configure the at least one network parameter using the recommended at least one network parameter value; and- initiating (206) configuration of the recommended at least one network parameter value when it has been determined to configure the at least one network parameter using the recommended at least one network parameter value and requesting the first application (106a) to recommend at least one alternative network parameter value when it has been determined not to configure the at least one network parameter using the recommended at least one network parameter value.

2. The method (200) according to claim 1, wherein the step (204) of determining whether or not to configure the at least one network parameter using the recommended at least one network parameter value comprises:- identifying (302) a plurality of Key Performance Indicator, KPI, metrics associated with the at least one network parameter for which the at least one network parameter value is recommended;- determining (304) at least one KPI metric from the plurality of KPI metrics of the at least one network parameter targeted by the recommended at least one network parameter value;- determining (306) whether there exists any conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value; and- when it has been determined that there is no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, determining (308) to configure the at least one network parameter using the recommended at least one network parameter value.

3. The method (200) according to claim 2, wherein when it has been determined that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, determining (310) not to configure the at least one network parameter using the recommended at least one network parameter value.

4. The method (200) according to any of claims 2-3, wherein the step (306) of determining whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value comprises:- selecting, from lookup tables, LUTs, built for a number of possible states of the network node (114), a LUT for a current state of the network node (114) for which the at least one network parameter value is recommended; and- determining, using the selected LUT, whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value.

5. The method (200) according to claim 4, wherein the LUT built for the state of the network node (114) comprises an association value for each combination of a network parameter and a respective KPI metric, wherein the association value indicates an association of the KPI metric with the respective network parameter.

6. The method (200) according to any of claims 4-5, wherein the step of selecting the LUT for the current state of the network node (114) is preceded by the following steps:- identifying the number of possible states of the network node (114), wherein each possible state representing characteristics of the network node (114) in terms of a plurality of network parameters and KPI metrics associated with each of the plurality of network parameters; and- building the LUTs for the number of possible states of the network node (114).

7. The method (200) according to claim 6, wherein the step of identifying the number of possible states of the network node (114) comprising:- determining a maximum possible number of states of the network node (114) by evaluating each of the plurality of network parameters of the network node (114) and respective KPI metrics using a first learning model (716);- building the LUT for each state of the determined maximum possible number of states; and- grouping the states of the network node (114) when the LUTs associated with said states of the network node (114) differ from each other by a pre-defined value, wherein the grouped states are identified as the number of possible states of the network node (114).

8. The method (200) according to any of claims 6-7, wherein the step of building the LUT for a state of the identified number of possible states of the network node (114) comprises:- obtaining historical data collected for a pre-determined time interval, said historical data comprising a plurality of network parameter values recommended by the plurality of applications (106a-106n), a state of the network node (114) associated with at least one of the recommended plurality of network parameter values, at least one network parameter configured using at least one of the plurality of network parameter values, and the KPI metrics resultant from the configuration of said at least one network parameter; anddetermining the association value for each combination of one of the plurality of network parameters and one of respective KPI metrics, by evaluating the historical data using a second learning model (718).

9. The method (200) according to claim 8, wherein the second learning model (718) evaluates the historical data based on at least one of:- identification of at least one feature as an importance from the obtained historical data;- interpolation between the states of the network node (114), when the states have been arranged in an order;- similarity measures between the states of the network node (114), when the states have not been arranged in the order; and- an input received from a radio domain expert.

10. The method (200) according to any of claims 4-9, wherein the step of selecting the LUT for the current state of the network node (114) for which the at least one network parameter value is recommended comprises:- detecting, using a distance metric, whetherany of the identified number of states of the network node (114) matches with the current state of the network node (114) for which the at least one network parameter value is recommended; and- when it has been determined that one of the identified number of states of the network node (114) matches with the current state of the network node (114), selecting the LUT corresponding to said state as the LUT for the current state of the network node (114).

11. The method (200) according to any of claims 4-10, wherein the step (408) of determining, using the selected LUT, whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value comprises:- determining, using the selected LUT, an weighted sum of association value of the recommended at least one network parameter value with respect to each of theother KPI metrics different from the targeted KPI metric of the at least one network parameter;- comparing the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics with a threshold factor;- determining that there exists no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value when the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics is smaller than the threshold factor; and- determining that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value when the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics is greater than the threshold factor.

12. The method (200) according to claim 11, wherein the step of determining, using the selected LUT, the weighted sum of association value of the recommended at least one parameter value with respect to each of the other KPI metrics different from the targeted KPI metric of the at least one network parameter comprises:- identifying from the selected LUT a priority and an association value associated with each of the other KPI metrics with respect to the at least one network parameter;- determining weighted results for the other KPI metrics based on the priority and the association value of each of the other KPI metrics; and- summing up the weighted results to determine the weighted sum of association value of the recommended at least one parameter value.

13. The method (200) according to claim 11, further comprising a step of tuning the threshold factor, said step comprising: evaluating the historical data using a third learning model (720) to determine the at least one network parameter configured with a minimum configuration timeand a minimum impact on other KPI metrics different from the targeted KPI metric of the at least one network parameter; and tuning the threshold factor based on the evaluation.

14. The method (200) according to any of claims 4-13, further comprising a step of updating the LUT, said step comprising:- receiving update information related to at least one application, wherein the update information indicates at least one of: on-boarding of a new application, and a subscription of the at least one application targeting new KPI metrics; and- updating the LUT in accordance with the received update information.

15. The method (200) according to any of the preceding claims, wherein the step (204) of determining whether or not to configure the at least one network parameter using the recommended at least one network parameter value comprises:- determining (402) an weighted sum of association value of the recommended at least one network parameter value using a LUT selected for the current state of the network node (114);- comparing (404) the weighted sum of association value of the at least one network parameter value with an weighted sum of association value of at least one other network parameter value recommended by a second application (106b) among the plurality of applications (106a-106n) for the same at least one network parameter;- determining (406) to configure the at least one network parameter using the recommended at least one network parameter value, if the weighted sum of association value of the at least one network parameter value is smaller than the weighted sum of association value of the at least one other network parameter value recommended by the second application (106b); and- determining (406) not to configure the at least one network parameter using the recommended at least one network parameter value, if the weighted sum of association value of the at least one network parameter value is greater than theweighted sum of association value of the at least one other network parameter value recommended by the second application (106b).

16. The method (200) according to any of the preceding claims, wherein the step (206) of initiating the configuration of the at least one network parameter value comprises:- adjusting the at least one network parameter using the recommended at least one network parameter value.

17. The method (200) according to any of the preceding claims, wherein the step (206) of requesting the first application (106a) for at least one alternate network configuration comprises:- requesting the first application (106a) to reduce a search space of the at least one network parameter value.

18. An apparatus of a network node (114) configured for handling network parameter recommendations from a plurality of applications (106a-106n), said plurality of applications (106a-106n) being executed on a non-real time Radio Access Network, RAN, Intelligent Controller, RIC (108), of an automation platform (102), the network node (114) being in communication with the automation platform (102), the apparatus comprising a controlling circuitry (706) configured to cause:- reception of a network parameter recommendation from a first application (106a) among the plurality of applications (106a-106n), said network parameter recommendation comprising at least one recommended network parameter value for configuring at least one network parameter of the network node (114);- determination of whether or not to configure the at least one network parameter using the recommended at least one network parameter value; and- initiation of configuration of the at least one recommended network parameter value when it has been determined to configure the at least one network parameter using the recommended at least one network parameter value and requesting of the first application (106a) to recommend at least one alternative network parameter value when it has been determined not to configure the atleast one network parameter using the recommended at least one network parameter value.

19. The apparatus according to claim 18, wherein the controlling circuitry (706) is configured to cause determination of whether or not to configure the at least one network parameter using the recommended at least one network parameter value by causing:- identification of a plurality of Key Performance Indicator, KPI, metrics associated with the at least one network parameter for which the at least one network parameter value is recommended;- determination of at least one KPI metric from the plurality of KPI metrics of the at least one network parameter targeted by the recommended at least one network parameter value;- determination of whether there exists any conflict with other KPI metrics different from the targeted at least one KPI metric of the at least one network parameter from the recommended at least one network parameter value; and- when it has been determined that there is no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, determination to configure the at least one network parameter using the recommended at least one network parameter value.

20. The apparatus according to claim 19, wherein when it has been determined that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value, the controlling circuitry (706) is configured to cause determination of not to configure the at least one network parameter using the recommended at least one network parameter value.

21. The apparatus according to any of claims 19-20, wherein the controlling circuitry (706) is configured to cause determination of whetherthere exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value by causing:- selection of a lookup table, LUT, from LUTs built for a number of possible states of the network node (114), for a current state of the network node (114) for which the at least one network parameter value is recommended; and- determination of, using the selected LUT (602), whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value.

22. The apparatus according to claim 21, wherein the LUT (602) built for the state of the network node (114) comprises an association value for each combination of a network parameter and a respective KPI metric, wherein the association value indicates an association of the KPI metric with the respective network parameter.

23. The apparatus according to any of claims 21-22, wherein the controlling circuitry (706) is configured to cause selection of the LUT for the current state of the network node (114) that is preceded by causing:- identification of the number of possible states of the network node (114), wherein each possible state representing characteristics of the network node (114) in terms of a plurality of network parameters and KPI metrics associated with each of the plurality of network parameters; and- building of the LUTs forthe number of possible states of the network node (114).

24. The apparatus according to claim 23, wherein the controlling circuitry is configured to cause identification of the number of possible states of the network node (114) by causing:- determining a maximum possible number of states of the network node (114) by evaluating each of the plurality of network parameters of the network node (114) and respective KPI metrics using a first learning model (716);- building the LUT for each state of the determined maximum possible number of states; and- grouping the states of the network node (114) when the LUTs associated with said states of the network node (114) differ from each other by a pre-definedvalue, wherein the grouped states are identified as the number of possible states of the network node (114).

25. The apparatus according to any of claims 23-24, wherein the controlling circuitry (706) is configured to cause building of the LUT (602) for a state of the identified number of possible states of the network node (114) by causing:- obtaining of historical data collected for a pre-determined time interval, said historical data comprising a plurality of network parameter values recommended by the plurality of applications (106a-106n), a state of the network node (114) associated with at least one of the recommended plurality of network parameter values, at least one network parameter configured using at least one of the plurality of network parameter values, and the KPI metrics resultant from the configuration of said at least one network parameter; and- determination of the association value for each combination of one of the plurality of network parameters and one of respective KPI metrics, by evaluating the historical data using a second learning model (718).

26. The apparatus according to claim 25, wherein the second learning model (718) evaluates the historical data to determine the association value based on at least one of:- identification of at least one feature as an importance from the obtained historical data;- interpolation between the states of the network node (114), when the states have been arranged in an order;- similarity measures between the states of the network node (114), when the states have not been arranged in the order; and- an input received from a radio domain expert.

27. The apparatus according to any of claims 21-26, wherein the controlling circuitry (706) is configured to cause selection of the LUT for the current state of the network node (114) for which the at least one network parameter value is recommended by causing:- determination of, using a distance metric, whether any of the identified number of states of the network node (114) matches with the current state of the network node (114) for which the at least one network parameter value is recommended; and- when it has been determined that one of the identified number of states of the network node (114) matches with the current state of the network node (114), selection of the LUT corresponding to said state as the LUT for the current state of the network node (114).

28. The apparatus according to any of claims 21-27, wherein the controlling circuitry (706) is configured to cause determination of, using the selected LUT, whether there exists any conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value by causing:- determination of, using the selected LUT, an weighted sum of association value of the recommended at least one network parameter value with respect to each of the other KPI metrics different from the targeted KPI metric of the at least one network parameter;- comparison of the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics with a threshold factor;- determination that there exists no conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value when the weighted sum of association value of the at least one network parameter value for each of the other KPI metrics is smaller than the threshold factor; and- determination that there exists a conflict with the other KPI metrics of the at least one network parameter from the recommended at least one network parameter value when the weighted sum of association value of the at least one network parameter value is greater than the threshold factor.

29. The apparatus according to claim 28, wherein the controlling circuitry (706) is configured to cause determination of, using the selected LUT, the weighted sum ofassociation value of the recommended at least one parameter value with respect to each of the other KPI metrics different from the targeted KPI metric of the at least one network parameter by causing:- identification from the selected LUT a priority and an association value associated with each of the other KPI metrics with respect to the at least one network parameter;- determination of weighted results for the other KPI metrics based on the priority and the association value of each of the other KPI metrics; and- summing up of the weighted results to determine the weighted sum of association value of the recommended at least one parameter value.

30. The apparatus according to claim 28, wherein the controlling circuitry (706) is further configured to cause tuning of the threshold factor by causing:- evaluation of the historical data using a third learning model (720) to determine the at least one network parameter configured with a minimum configuration time and a minimum impact on other KPI metrics different from the targeted KPI metric of the at least one network parameter; and- tuning of the threshold factor based on the evaluation.

31. The apparatus according to any of claims 21-30, wherein the controlling circuitry (706) is further configured to cause updating of the LUT by causing:- reception of update information related to at least one application, wherein the update information indicates at least one of: on-boarding of a new application, and a subscription of the at least one application targeting new KPI metrics; and- updating of the LUT in accordance with the received update information.

32. The apparatus according to any of claims 18-31, wherein the controlling circuitry (706) is configured to cause determination of whether or not to configure the at least one network parameter using the recommended at least one network parameter value by causing:- determination of an weighted sum of association value of the recommended at least one network parameter value using a LUT selected for the current state of the network node (114);- comparison of the weighted sum of association value of the at least one network parameter value with an weighted sum of association value of at least one other network parameter value recommended by a second application (106b) among the plurality of applications (106a-106n) for the same at least one network parameter;- determination to configure the at least one network parameter using the recommended at least one network parameter value, if the weighted sum of association value of the at least one network parameter value is smaller than the weighted sum of association value of the at least one other network parameter value recommended by the second application (106b); and- determination not to configure the at least one network parameter using the recommended at least one network parameter value, if the weighted sum of association value of the at least one network parameter value is smaller than the weighted sum of association value of the at least one other network parameter value recommended by the second application (106b).

33. The apparatus according to any of claims 18-32, wherein the controlling circuitry (706) is configured to cause initiation of the configuration of the at least one network parameter value by causing:- adjusting of the at least one network parameter using the recommended at least one network parameter value.

34. The apparatus according to any of claims 18-33, wherein the controlling circuitry (706) is configured to cause requesting the first application (106a) to recommend the at least one alternative network parameter value by causing: requesting of the first application (106a) to reduce a search space of the at least one network parameter value.

35. A network node (114) comprising the apparatus of any of the claims 18 through 34.

36. A computer program product comprising a non-transitory computer readable medium, having thereon a computer program comprising program instructions, the computer program is loadable into a data processing unit and configured to cause execution of the method according to any of claims 1 through 17 when the computer program is run by the data processing unit.

Images