Method and system for experimentally assessing the effectiveness of optimization actions on a mobile communications network

Abstract

A computer-implemented method, and a system configured for implementing the method, for assessing the change in the Quality of Service, QoS, offered by a mobile communications network, consequent to a change in configuration parameters of one or more network cells of the mobile communications network. A network cell cluster is defined, including the one or more network cells and neighboring network cells, in such a way as to correspond to a mobile communications network coverage area that remains essentially constant before and after the change in configuration parameters of the one or more network cells. Network measurements are acquired in respect of an area of the mobile communications network corresponding to the network cell cluster. During a first acquisition time interval before the configuration parameters change, first network traffic amount indicators and first QoS indicators are acquired, and during a second acquisition time interval after the configuration parameters change, second network traffic amount indicators and second QoS indicators are acquired. From the acquired first network traffic amount indicators a first distribution of network traffic amount during the first acquisition time interval is derived, and from the acquired second network traffic amount indicators a second distribution of network traffic amount during the second acquisition time interval is derived. A reference distribution of network traffic amount in the network cell cluster during a reference time interval is obtained. A first reward based on the first QoS indicators and a second reward based on the second QoS indicators are calculated. The change in the QoS is assessed based on a comparison between the first and the second rewards. For calculating the first and the second rewards, first weights are calculated, to be applied to the first QoS indicators, and second weights are calculated, to be applied to the second QoS indicators. The first and second weights are calculated so as to make the first distribution of network traffic amount and the second distribution of network traffic amount correspond to the reference distribution of network traffic amount.

Description

[0001]METHOD AND SYSTEM FOR EXPERIMENTALLY ASSESSING THE EFFECTIVENESS OF OPTIMIZATION ACTIONS ON A MOBILE COMMUNICATIONS NETWORK ** * * * *DESCRIPTION Technical field The present disclosure relates to the field of mobile communications networks. In particular, the present disclosure relates to the optimization of mobile radiocommunications networks, particularly but not limitatively 5G or future generationsnetworks. Specifically, the present disclosure relates to a method, and a relatedsystem, for assessing the effectiveness of optimization actions on mobile radiocommunications networks (mobile radio networks).Technical background With a view to optimizing the Quality of Service (QoS) experienced by usersin mobile radio networks, with particular reference to user throughput (the amount ofdata received – correctly received bits - by a user of user equipment within a networkcell over certain period of time), it is useful to compensate for asymmetries in thedistribution of network traffic on the mobile radio network cells. Such asymmetries in the network traffic distribution can be compensated by acting on modifiableparameters of the Radio Access Network (RAN). Actions on the RAN modifiableparameters include actions on network cells’ configuration parameters, so as to obtainthe right compromise between the Signal-to-Interference-plus-Noise Ratio (SINR) anduser traffic distribution per network cell. Network cells include radio transceiver stations comprising one or moreantennas. Optimization actions on a network cell’s configuration parameters mayinvolve an optimization of the network cell’s antenna parameters (transmission power,electrical tilt and azimuth of the antennas, and any other available parameter affecting radiation diagram and power spatial distribution, including parameters controllingradiation patterns for active antennas, for beamforming techniques, which are typicalof 5G networks), as well as an optimization of those parameters affecting network cellselection procedures and handover, in order to make the traffic volumes and thenumber of served users per cell (traffic distribution in the cells) as homogeneous as possible. User throughput depends on the SINR guaranteed by the mobile radio networkin a specific pixel (one of the elementary territory units into which a territory coveredby a mobile radio network is conventionally subdivided), but it is also inversely proportional to the number of active users in each individual network cell. An actionon a network cell’s coverage area, carried out through modifications of the networkcell’s antenna(s) pointing and / or of the transmission power and / or of the parametersof the network cell selection and handover procedure, can therefore allow a better distribution of users between cells, although at the potential expense of optimal planning of the offered SINR, in order to avoid critical areas in terms of low userthroughput due to strong concentrations of users in individual network cells. The samegoes for stability in terms of latency, which increases as the number of users increases. In fact, a more homogeneous distribution of users in the various network cellsallows avoiding areas of greater QoS degradation, in terms of user throughput.In network planning, there is a tendency to try to improve the most critical areas, for example by maximizing the user throughput of the worst fifth (5th) percentile(generally speaking, the 5th percentile is a value of a certain parameter – the userthroughput in the case here considered – at which only the 5% of the consideredpopulation – the users of user equipment in this case – experiences lower values ofsaid parameter, i.e., lower values of user throughput) referred to more congested areas,since in areas with good QoS the user throughput worsening is not perceived by users as critically as it is instead perceived by users in congested areas. At the same time, optimization can include traffic-based shutdown of networksites (base stations) to reduce energy consumption and therefore operating costs forthe Telco. The right compromise between QoS and energy consumption can be anobjective of the network optimization. In order to redistribute user traffic across the network cells, two differentnetwork solutions can be adopted: Coverage and Capacity Optimization (CCO), whichenvisages actions on network cells’ antenna parameters such as pointing and power -and Mobile Load Balancing (MLB), which operates on parameters of the cell selection and handover procedures. In order to proceed with an optimization of the RAN parameters, a map of thetraffic sources in the territory is necessary. Such a traffic sources map can be obtainedwith Minimization of Drive Test (MDT) traffic source profiles, or from trafficmeasurements per cell mapped onto pixels, using information on building density androad conditions in the territory covered by the mobile radio network.Exploiting this traffic sources distribution map across the territory, the optimalconfiguration of the mobile radio network can be estimated for generic traffic profileand user profile considered.The network optimization based on MLB and network sites switch-on / switch-off can be static, i.e., an optimal fixed network configuration is calculated for areference distribution of traffic sources, or dynamic, i.e., a different networkconfiguration can be foreseen: a network configuration for peak time(s), withasymmetry of users which tends to rebalance between the cells, and a network configuration for off-peak times, in which the network configuration tends tomaximize network capacity, since in off-peak times there are no critical issues, due tothe low number of users and their homogeneous distribution. The network will choose one or the other network configuration based on the user traffic profile of one or more cells. A redistribution of users across the network cells may be detrimental in termsof network capacity but, as mentioned, it may bring about a mitigation of significantcritical issues in certain areas, improving the QoS that the mobile network provides tothe most penalized users, the criterion on which the network is planned. The action of modifying RAN parameters for the purpose of networkoptimization is studied and defined with simulation tools. For example, WO 2022 / 207402 discloses a method, implemented by a dataprocessing system, of adjusting modifiable parameters of network cells of a self- organizing cellular mobile communications network, comprising: retrieving a current configuration of network cells currently deployed on field, including modifiable parameters; exploring different configurations of the network cells, each differing from the retrieved current configuration and from other different configurations by a change in the value of at least one of the modifiable parameters of at least one network cell; evaluating the explored different configurations of the network cells. When a suitable degree of goodness is assessed for a certain configuration of network cells, said new configuration of network cells is automatically deployed on field by modifying one or more of the modifiable parameters of network cells. Another example is provided in Marco Skocaj et al., “Cellular Network Capacity and Coverage Enhancement with MDT Data and Deep Reinforcement Learning”, Computer Communications Volume 195, 1 November 2022, Pages 403- 415. The authors investigate a MDT-driven Deep Reinforcement Learning (DRL) algorithm to optimize coverage and capacity by tuning antennas tilts on a cluster ofcells from a cellular network of an Italian Telco. MDT data, electromagneticsimulations, and network Key Performance indicators (KPIs) are jointly utilized to define a simulated network environment for the training of a Deep Q-Network (DQN) agent. EP 4175370 describes that to maximize power saving in a RAN comprisingcells, using a trained model an optimal action amongst actions comprising switching on one or more cells, switching off one or more cells, and doing nothing is determined,which maximizes a long-term reward on tradeoff between throughput and power. Thetrained model takes as input a load estimate. The trained model may be updated onlineusing measurement results on load, throughput and power consumption. Summary As mentioned before, the action of modifying RAN parameters for the purposeof mobile radio network optimization is studied and defined with simulation tools. From tests carried out on the mobile radio network in operation on the field, itis relatively easy to carry out analytical evaluations to define the best networkconfiguration. However, the Applicant has understood that the real network operating conditions may not be correctly considered in the design process, due to the partial randomness of some aspects. Among these aspects there are, in the context of the evaluation of simulated and / or real network performances, the distribution of users over the territory, their mobility conditions and pathloss profiles: in the networkoptimization process an experimental validation of the proper modelling of theseeffects is therefore necessary. The Applicant has perceived that an on-field assessment of the effectiveness ofthe design process, and of the changes in the network configuration applied to themobile radio network, is necessary. This assessment can be done after havingimplemented on the field the network configuration change, by analyzing statisticalcounters and comparing them with those acquired before the implementation of thenetwork optimization. The Applicant has observed that known solutions, like those disclosed in WO2022 / 207402, in the paper by Marco Skocaj et al., and in EP 4175370, are notconcerned with such an experimental verification / validation of the effectiveness of the implemented network configuration change. An aim of such a verification should be to demonstrate that the optimizationphase has achieved the expected results, particularly in terms of QoS improvement, for example consequent to an improvement in user throughput. The Applicant has observed that when network configuration changes aredeployed to the mobile radio network on the field, it is particularly challenging todefine a metric to evaluate the positive effects, or rather the ineffectiveness, of thedeployed optimized network configuration on the QoS, particularly on the userthroughput. In particular, the distribution of users (and therefore of traffic sources)across the territory is the most evident and important source of uncertainty present inthe optimization process procedure. It is in fact not easy to compare, by analyzing thenetwork statistical counters, the QoS that the mobile network is able to provide to theusers in operating conditions before and after the network configuration change,because the distribution profiles of the offered traffic across the territory changesignificantly and the distribution of the traffic sources has an impact directly on detected user throughput, which, as mentioned, plays an important role in the quantification of the QoS. An object of the present disclosure is to define a methodology to mitigate the effects of changes in the distribution of users and traffic volumes across the territory covered by a mobile radio network, in order to be able to carry out a verificationprocess capable of reliably assessing the effectiveness of the deployed networkconfiguration change in normal operating conditions of the mobile radio network. Essentially, the present disclosure proposes a method, and a related system,capable of making the verification process independent of these variations in thedistribution of users and traffic volumes across the territory, thanks to the introduction- in the comparison between experimental data collected before and after theimplementation, in the mobile network, of the network configuration modification - ofappropriate criteria for homogenizing the distribution of users, and thus of network traffic, before and after the implementation of the network configuration change. Inthis way, it is possible to guarantee a correct modelling of pathloss, mobility and othermodelling aspects, regardless of the distribution of users.According to an aspect of the present invention, a computer-implemented method is proposed for assessing the change in the Quality of Service, QoS, offered by a mobile communications network, consequent to a change in configuration parameters of one or more network cells of the mobile communications network. The method comprises defining a network cell cluster including the one ormore network cells and neighboring network cells, adjacent to the one or more networkcells. Advantageously, the network cell cluster is defined in such a way as tocorrespond to a mobile communications network coverage area that remainsessentially constant before and after the change in configuration parameters of the oneor more network cells. The method comprises acquiring network measurements in respect of an area of the mobile communications network corresponding to the network cell cluster.Acquiring network measurements comprises: acquiring, during a first acquisition timeinterval before said change in configuration parameters, first network traffic amount indicators and first QoS indicators, and acquiring, during a second acquisition time interval after said change in configuration parameters, second network traffic amount indicators and second QoS indicators. The method comprises: deriving from the acquired first network traffic amount indicators a first distribution of network traffic amount during the first acquisition time interval, and deriving from the acquired second network traffic amount indicators asecond distribution of network traffic amount during the second acquisition timeinterval. Areference distribution of network traffic amount is obtained, adapted toprovide a reference distribution of network traffic amount in the network cell cluster during a reference time interval. The method further comprises: calculating a first reward based on said first QoS indicators and calculating a second reward based on said second QoS indicators, and assessing the change in the QoS based on a comparison between the first reward and the second reward. Calculating the first reward and the second reward comprises calculating firstweights to be applied to the first QoS indicators and second weights to be applied to the second QoS indicators, said first weights and second weights being calculated soas to make the first distribution of network traffic amount and the second distributionof network traffic amount correspond to the reference distribution of network traffic amount. The network traffic amount indicators may be any one of, or one or more of: indicators of network traffic volume (e.g., expressed in kbits), indicators of userpresences in the network cells (i.e., users for which the mobile network records a lastevent in the network cells, derived from indications about the location of the user whenthe last event has been recorded), indicators of numbers of active users in the networkcells. The first and second QoS indicators may include indicators of a distribution of network traffic amounts as a function of a user throughput that the mobile communications network is capable of providing to users. In embodiments of the present invention, the first acquisition time intervalcomprises a first succession of first elementary acquisition time intervals, and thesecond acquisition time interval comprises a second succession of second elementary acquisition time intervals. The first network traffic amount indicators and first QoSindicators are acquired during each first elementary acquisition time interval of thefirst succession, and the second network traffic amount indicators and second QoSindicators are acquired during each second elementary acquisition time interval of thesecond succession. In embodiments of the present invention, the first distribution of network trafficamount during the first acquisition time interval is a distribution of the first elementaryacquisition time intervals as a function of the first network traffic amount indicators,and the second distribution of network traffic amount during the second acquisitiontime interval is a distribution of the second elementary acquisition time intervals as a function of the first network traffic amount indicators. In embodiments of the present invention, obtaining a reference distribution ofnetwork traffic amount comprises:- defining a reference acquisition time interval, wherein said referenceacquisition time interval comprises a third succession of third elementary acquisitiontime intervals; -acquiring, during each third elementary acquisition time interval of the thirdsuccession, third network traffic amount indicators and third QoS indicators, and- deriving from the third network traffic amount indicators the referencedistribution of network traffic amount, the reference distribution of network traffic amount being a distribution of the third elementary acquisition time intervals as afunction of the third network traffic amount indicators.In other embodiments of the present invention, the reference distribution ofnetwork traffic amount is the first distribution of network traffic amount during the first acquisition time interval, and obtaining a reference distribution of network trafficamount includes deriving from the first network traffic amount indicators the firstdistribution of network traffic amount during the first acquisition time interval. Thefirst weights to be applied to the first QoS indicators are unitary weights and the second weights are calculated so as to make the second distribution of network traffic amount correspond to the first distribution of network traffic amount. In embodiments of the present invention, the weights are calculated so as toobtain, from the distribution of the second elementary acquisition time intervals as a function of the first network traffic amount indicators, a re-shaped distribution having a shape corresponding to one among: the distribution of the third elementaryacquisition time intervals as a function of the third network traffic amount indicators,the distribution of the first elementary acquisition time intervals as a function of the first network traffic amount indicators. In embodiments of the present invention, calculating weights to be applied to the second network measurements comprises: -defining a plurality of classes of network traffic amount in respect of thereference network traffic amount distribution, each class of network traffic amount ofsaid plurality of classes of network traffic amount corresponding to a respective reference network traffic amount; -classifying each of the first elementary acquisition time intervals and each ofthe second elementary acquisition time intervals in a respective class of said pluralityof classes according to the respective first and second network traffic amount indicators; -determining a number of first elementary acquisition time intervals andsecond elementary acquisition time intervals in each class of said plurality of classes,and -calculating the weights to be applied to the second network measurementscomprises based on the determined number of first elementary acquisition timeintervals and second elementary acquisition time intervals in each class of saidplurality of classes. In embodiments of the present invention, the method comprises aggregating,over all the network cells of the cell cluster, the acquired first network traffic amount indicators and first QoS indicators, and the acquired second network traffic amount indicators and second QoS indicators.In embodiments of the present invention, the first and second network trafficamount indicators, and the first and second QoS indicators are acquired from themobile communications network as data aggregated at the level of each network cell of the cell cluster. In other embodiments of the present invention, the first and second networktraffic amount indicators, and the first and second QoS indicators are acquired as geo-referenced data, particularly MDT data. The method may further comprise aggregating the acquired geo-referenced data at the level of sub-areas of the coverage area of each network cell of the cell cluster. According to another aspect of the present invention, a system is proposed for assessing the change in the Quality of Service, QoS, offered by a mobile communications network, consequent to a change in configuration parameters of one or more network cells of the mobile communications network. The system is configured to acquire network measurements in respect of anarea of the mobile communications network corresponding to a network cell clusterincluding the one or more network cells and neighboring network cells adjacent to theone or more network cells. Advantageously, the network cell cluster corresponds to amobile communications network coverage area that remains essentially constant before and after the change in configuration parameters of the one or more network cells. The system is configured to acquire said network measurements by: acquiring,during a first acquisition time interval before said change in configuration parameters,first network traffic amount indicators and first QoS indicators, and acquiring, duringa second acquisition time interval after said change in configuration parameters,second network traffic amount indicators and second QoS indicators.The system is also configured to obtain a reference distribution of networktraffic amount adapted to provide a reference distribution of network traffic amount in the network cell cluster during a reference time interval. The system is further configured to: -derive from the first network traffic amount indicators a first distribution of network traffic amount during the first acquisition time interval; -derive from the acquired second network traffic amount indicators a seconddistribution of network traffic amount during the second acquisition time interval; -calculate a first reward based on said first QoS indicators and calculating asecond reward based on said second QoS indicators; -assess the change in the QoS based on a comparison between the first rewardand the second reward. The system is configured to calculate the first reward and the second reward bycalculating first weights to be applied to the first QoS indicators and second weightsto be applied to the second QoS indicators, the first weights and the second weightsbeing calculated so as to make the first distribution of network traffic amount and thesecond distribution of network traffic amount correspond to the reference distribution of network traffic amount. The method and system of the present disclosure are aimed at experimentally validating mobile network optimization actions involving changes in the configuration of the mobile network on the field. An optimization action, i.e., a change in the mobile network configuration, that is experimentally assessed to bring about an improvement to the QoS is retained, while an optimization action that does not bring about a significant QoS improvement (not to say a worsening of the QoS) may be discarded. After the implementation, in the mobile network, of a change in the network configuration (e.g., changes in antennas’ electrical tilts and / or azimuth, transmitted power, network cells’ selection / re-selection thresholds), the values of the networkperformance indicators (KPIs) that, prior to the change of the network configuration,were unsatisfactory and led to the execution of the network optimization action, are assessed: if the new KPIs values are worse than before (or possibly even if the new KPIs values are not significantly better than before), the configuration of the mobile network may be prudentially switched back to the previous network configuration. As known, the network optimization process is at least partly based onsimulations (being impractical to continuously act on the mobile network just to try and see if a certain change in the network configuration works or not). Thanks to the method and system of the present disclosure, if it is experimentally assessed that the change in the QoS after a change in the network configuration is not satisfactory (because the QoS after the change has not significantly changed, not to say the QoS after the change is worse than before), then, one or more of the following actions may be undertaken. One possible action is to re-run the optimization / simulation algorithm by modifying: -the ranges of variability of the modifiable parameters (e.g., antennas’electrical tilts and / or azimuth, transmitted power, network cells’ selection / re-selection thresholds), so as to increase the degrees of freedom of the networkoptimization / simulation algorithm;- the set of modifiable parameters that the network optimization / simulationalgorithm can consider (if for example in the previous run of theoptimization / simulation algorithm not all the possible modifiable parameters have been taken into consideration); -the set of network cells whose parameters can be modified;- the importance (weights) assigned to the different items (e.g., traffic volume,user throughput) considered in the calculation of the reward / cost function for thenetwork configurations before and after the change, thereby modifying the goals set to the optimization / simulation algorithm; -as a last resource, modifying the optimization / simulation algorithm (forexample, in case of optimization / simulation algorithms based on machine learning / artificial intelligence, this may involve a re-training of the algorithm). Brief description of the drawings Features and advantages of the solution here disclosed, including those mentioned in the foregoing, will appear more clearly by reading the following detailed description of exemplary and non-limitative embodiments. For a better intelligibility, the following description should be read making reference to the annexed drawings, wherein:- Fig. 1 is a pictorial view of an exemplary geographic area of interest coveredby network cells of a mobile radio network involved in a network optimization procedure; -Fig. 2 summarizes, in the form of an activity diagram, the activities of amethod according to an embodiment of the present invention; -Fig. 3 shows first and second acquisition time intervals during whichexperimental data (network measurements and Key Performance Indicators - KPIs)are acquired, before and after an implementation of a network configuration change in the mobile radio network of Fig.1; -Fig. 4 schematizes a generic elementary acquisition time interval (eachelementary acquisition time interval being a “Record Of Period”, shortly “ROP”) of the first and second acquisition time intervals of Fig.3; -Fig. 5 shows an exemplary distribution of traffic volume (TrV in kB, inordinate) as a function of the user throughput (ThrP in kbit / s, in abscissa);- Fig. 6A depicts a classification of the measurements acquired in theelementary acquisition time intervals (ROPs) before the implementation of the network configuration change in classes of number of users (traffic volume), and a corresponding traffic distribution profile; -Fig. 6B depicts a classification of the measurements acquired in theelementary acquisition time intervals (ROPs) after the implementation of the networkconfiguration change, in classes of traffic volume, and a corresponding trafficdistribution profile; -Fig. 6C depicts the classification of the measurements of Fig. 6B after aweighting action for making the traffic distribution profile in the acquisition time interval after the implementation of the network configuration change comparable with the traffic distribution profile in the acquisition time interval before the implementation of the network configuration change; -Fig. 7 schematically shows, in terms of functional blocks, some modules of asystem according to an embodiment of the present invention for assessing the effectiveness of a network configuration change; -Fig. 8 shows, in addition to the acquisition time intervals of Fig. 3, a referenceacquisition time interval during which experimental data (network measurements andKPIs) are acquired, in accordance with another embodiment of the present invention; -Fig. 9A depicts a classification in classes of number of users (traffic volume)of the measurements acquired in the elementary reference acquisition time intervals (ROPs) of the reference acquisition time interval of Fig.8; -Fig. 9B and Fig. 9C depict classifications in classes of traffic volume of themeasurements acquired in the elementary acquisition time intervals (ROPs) before and, respectively, after the implementation of the network configuration change in terms of traffic volume, and a corresponding traffic distribution profile; -Fig. 10 schematizes zones of the coverage area of a network cell before andafter a change in the network configuration, and -Fig. 11 schematizes a definition of classes for classifying geo-referencedmeasurements of network traffic (average number of users) before and after a network configuration change. Detailed description of exemplary embodiments The disclosure in this document proposes a method, and a related system, forassessing the effectiveness of network configuration changes deployed in a mobile radio network as a consequence of network optimization actions. The disclosed method and system can be applied in the context of theimprovement or optimization of the QoS of a mobile radio network in a geographicarea of interest covered by the mobile radio network. Making reference to Fig.1, a (portion of a) mobile radio network, particularlya cellular mobile network (in short, “mobile network”) is schematically depicted, beingglobally denoted as 100. The mobile network 100 comprises a plurality of cellular communicationequipment 110 or base transceiver stations (shortly, “base stations”), e.g., eNodeBs orgNodeBs, each providing radio coverage over a respective area portion 115 of ageographic area covered by the mobile network. In the exemplary, simplified scenarioherein considered, each base station 110 is associated with a respective network cell,which represents the radio coverage area portion 115 of the base station, however, inpractical scenarios, each base station 110 may be associated with a plurality of networkcells, such as three network cells. According to an embodiment of the present invention, as illustrated by way ofexample, each network cell 115 is hexagonal in shape. In practice, though, the shapeof the network cells may differ significantly from an ideal hexagonal shape, e.g. dueto geographical and / or propagation characteristics or constraints of the area where the cell is located. According to an embodiment of the present invention, each base station 110comprises one or more electronic apparatuses (not shown). Example of electronic apparatuses include, but are not limited to, transceivers and digital signal processors.According to an embodiment of the present invention, each base station 110 comprisesone or more antennas. According to an embodiment of the present invention, a base station 110 allowsuser equipment UE of users located within the respective network cell 115 (andconnecting / connected to the mobile network 100) to exchange data traffic (e.g., webbrowsing, e-mailing, voice, or multimedia data traffic). The user equipment UE mayfor example comprise personal communication devices owned by users of the mobilenetwork 100 (the users being for example subscribers of services offered by the mobilenetwork 100). Examples of user equipment UE comprise, but are not limited to, mobile phones, smartphones, tablets, personal digital assistants and computers. According to an embodiment of the present invention, the base stations 110and their corresponding network cells 115 are part of a Radio Access Network (RAN)of the mobile network 100. In the following, a base station 110 and its correspondingnetwork cell 115 will be shortly referred to as a network cell 115.The RAN may be based on any suitable Radio Access Technology (RAT).Examples of RATs include, but are not limited to, UTRA (UMTS Terrestrial Radio Access), WCDMA (Wideband Code Division Multiple Access), CDMA2000, LTE (Long Term Evolution), LTE-A (LTE-Advanced), and NR (New Radio). According to an embodiment of the present invention, the RAN is communicably coupled with one or more core networks, such as the core network 120.The core network 120 may be any type of network configured to provide aggregation,authentication, call control / switching, charging, service invocation, gateway and subscriber database functionalities, or at least a subset (i.e., one or more) thereof. According to an embodiment of the present invention, the core network 120comprises a 4G / LTE core network, or a 5G core network, or a 6G core network.According to an embodiment of the present invention, the core network 120 iscommunicably coupled with other communication and / or data networks, such as the Internet and / or public switched telephone networks (not shown). The mobile network 100 may in particular be a Self-Organizing Network(SON). The mobile radio network may for example be a 5G mobile network or a 6G mobile network or a future generation mobile network, adopting the SON paradigm. ASON implements SON functionalities 125 that enable performing a CCO –MLB of the mobile network 100, i.e. functionalities that allow setting (i.e., tuning oradjusting) one or more operative parameters of the network cells 115 (hereinafter, cellparameters) for maximizing network performances to mobile network users on the field. According to an embodiment of the present invention, the cell parameters of each network cell define a configuration of that network cell (hereinafter, cell configuration), and the cell configurations of the network cells of the mobile network define as a whole a configuration of the mobile network (hereinafter, mobile networkconfiguration or network configuration).According to an embodiment of the present invention, the cell parameters comprise, but are not limited to, one or more of antenna parameters and cell selection / handover parameters. Examples of antenna parameters include, but are not limitedto, transmitted power, antenna electrical tilt, azimuth, gain and antenna radiationpattern (e.g. pointing direction, directivity and width of one or more lobes of the patternof lobes exhibited by the antenna radiation pattern). Example of cell selection / handover parameters include, but are not limited to, layer priority, minimum signal level for layer, intra-layer cell offset. According to the present invention, and as described in detail later on, the SONfunctionalities include a system 127 (QoS Improvement Assessment) configured andoperable to perform a method of assessment of the effectiveness of changes in thenetwork configuration implemented in the mobile network 100 as a result of CCOand / or MLB procedures carried out by the SON functionalities. Reference numeral 105 denotes a specific geographic area being a portion ofthe territory covered by the mobile network 100, which is assumed to be consideredfor purposes of improvement or optimization of the QoS with MLB and CCO.According to an embodiment of the present invention, the network cells 115covering the considered geographic area 105 are conveniently divided into three sets,as schematized in Fig.1: -a first set of network cells (“Target cells”) 115a: the Target cells 115a arenetwork cells susceptible of changes in their configuration, i.e., susceptible ofhaving their cell parameters modified for optimization purposes (for example,one or more antenna parameters, e.g., antenna(s) electrical tilt, antenna(s) azimuth, transmission power, beamforming parameters, or cell selection / handover parameters). Preferably, the number of Target cells 115a consideredis limited in number to, e.g., a maximum of about twenty cells;- a second set of cells (“Adjacent cells” or “Corona-1 cells” or “Tier-1 cells”)115b, geographically contiguous, i.e., in the neighbourhood of and adjacent to the Target cells 115a; the configuration parameters of the Corona-1 cells 115bcannot be modified, or are not to be modified, but these network cells arenonetheless affected (due to radio signal interference) by the Target cells 115a and can affect the Target cells 115a (i.e., interfere with the radio signalsthereof), such that the change in configuration of the Target cells 115a mayimply a change in the useful coverage area (best server area) of the Corona-1cells 115b; the Corona-1 cells 115b are taken into account to considerboundary effects of the network configuration optimization actions attempted by the SON functionalities 125 for CCO and MLB;- in embodiments of the present invention, a third set of cells (“Corona-2 cells”or “Tier-2 cells”) 115c is also considered; the Corona-2 cells 115c are cells, inthe neighbourhood of the Corona-1 cells 115b, whose cell parameters cannotbe modified / are not to be modified, and that are not affected by the Targetcells 115a, i.e., the best server areas of the Corona-2 cells 115c do not undergosignificant variations when the configuration (i.e., cells parameters) of theTarget cells 115a varies. The Corona-2 cells 115c are used to appropriatelydelimit the overall best server area of the cells of the first and second sets (Target cells 115a and Corona-1 cells 115b).In embodiments of the present invention, a “cell cluster” 130 is defined toinclude the Target cells 115a and the Corona-1 cells 115b, considered globally. Inother embodiments of the present invention, the “cell cluster” can be defined toinclude, in addition to the Target cells 115a and the Corona-1 cells 115b, also theCorona-2 cells 115c. In either case, the overall coverage area of the cell cluster 130remains constant (at least in a good approximation) after a change in the mobile network configuration is implemented. By considering the cell cluster 130 as a whole, instead of the individual networkcells, it is possible to introduce, in the comparison between the experimental data (measurements, statistical network counters, KPI) collected before and after the implementation of the network configuration change, a constraint on the mobile network carried traffic (and thus on the distribution of users) across the territory (whilesuch a constraint cannot be directly imposed on the carried traffic in individualnetwork cells, which changes, after the implementation of the network configurationchange, for the very reason that the network configuration change moves portions ofthe offered traffic from a network cell to another; later on, an alternative embodimentwill be described in which the cell cluster 130 is not considered as a whole).An overview of a method according to an embodiment of the present invention is provided below. The method according to an embodiment of the present invention includes thefollowing activities, actions or steps (not necessarily performed in the chronologicalorder in which they are presented below), schematically depicted in the activitydiagram of Fig. 2; reference is also made to the functional blocks diagram of Fig. 7.1. Activity 205Preliminarily, a network optimization method or algorithm to be used for theCCO and MLB of the mobile network 100 is identified and selected (e.g., by theTelco). Any one of the known network optimization algorithms can be chosen for thispurpose, for example the algorithm disclosed in WO 2022 / 207402 A1, or the algorithmpresented in the paper by Skocaj et al. “Cellular Network Capacity and Coverage Enhancement with MDT Data and Deep Reinforcement Learning”.2. Activity 210Still preliminarily, a metric is defined for the evaluation of mobile networkperformance, in particular a metric for the evaluation of the QoS experienced by themobile network users (QoS evaluation metric). As mentioned in the foregoing, in anembodiment of the present invention the user throughput (amount of data received –correctly received bits - by a user of a user equipment within a network cell over acertain period of time; for example, the user throughput can be expressed in kbits persecond, [kbits / s]) is taken in consideration in the evaluation of the QoS, thus the QoSevaluation metric is based on the user throughput. In particular, the QoS evaluation metric is based on the user throughput distribution, i.e., the spread of data transfer ratesexperienced by the users, or, otherwise stated, the distribution of the traffic offered tothe mobile network by the users as a function of the user throughput. For example, as will be described in greater detail later on, the QoS evaluation metric may be based on a properly weighted combination of the 5thpercentile of the user throughputdistribution and of the 50th percentile of the user throughput distribution.In embodiments of the present invention the QoS evaluation metric may also take into account, in addition to the user throughput, other parameters, for example parameters providing an indication of the number of handovers, and / or parameters providing an indication of how many base stations of the mobile network are “active”or “on”, i.e. ̧not “switched off”: indeed, the higher the number of “on” base stationsthe higher the user throughput, but this is achieved at the cost of a greater power consumption of the mobile network; the metric may then be defined in order to apply increasing penalties for increasing numbers of active base stations). In any case, thespecific metric defined for the evaluation of network performance does not constitutea limitation of the present invention.3. Activity 215During the operation of the mobile network 100, at a certain moment a situationemerges in the geographic area 105, and is detected (e.g., by the mobile networkmonitoring systems of the Telco), that makes a network optimization action (MLB andCCO) appropriate. For example, the detected situation may be an asymmetry in thedistribution of the offered traffic which persists in time, causing certain network cellsto be more heavily loaded than their neighbouring / adjacent cells; in this case there isa degradation in the mobile network performance, particularly in the experienced QoS in terms of user throughput, due to the presence of the congested network cells. The Telco may then decide to undertake a network optimization action (MLBand CCO), e.g., for distributing part of the traffic from the congested network cell toone or more adjacent network cells (despite the fact that such an action might negatively affect the whole mobile network capacity, a benefit may derive in terms of improved user throughput, and thus QoS). The selected optimization algorithm is thenexecuted and, as result, a modified network configuration, affecting the cell parametersof one or more of the Target cells 115a in the area 105, is identified whichimproves / optimizes the network performance; at a moment in time tc the identifiedmodified network configuration is deployed and implemented in the mobile network on field, modifying the cells parameters of one or more of the Target cells 115a.4. Activity 220 – Block 705As schematized in Fig. 3, two acquisition time intervals, INTA and INTB, aredefined, a first acquisition time interval INTA from instant tA1 to instant tA2 before thetime tc of implementation of the modified network configuration (networkconfiguration change) and a second acquisition time interval INTB from instant tB1 toinstant tB2 after the time tc of the network configuration change. Each one of the two acquisition time intervals INTA and INTB is relatively long, lasting for example some days or some weeks. It is pointed out that it is not relevant how much time before thetime tc of the network configuration change the first acquisition time interval INTA issituated, nor how much time after the time tc of the network configuration change thesecond acquisition time interval INTB is situated; the two acquisition time intervals INTAand INTBshould be chosen in such a way as to represent a generic condition of the mobile network before the network configuration change and, respectively, after the network configuration change (where by “generic condition” it is meant a conditionof the mobile network not biased by particular events, such as periods of national orlocal holidays, and / or time periods including events causing exceptional gatherings ofpeople in an area, like concerts or sport matches, and / or exceptional natural events).During the two acquisition time intervals, network measurements, statistical networkcounters and Key Performance Indicators (KPIs) are acquired. The KPIs which maybe useful to acquire during the two acquisition time intervals INTA and INTB include:number of users (Erlang), distribution of users, traffic volumes, throughput that themobile network is able to provide to the users (user throughput, particularly in uplink, or in downlink, or overall, i.e., in uplink and in downlink), distribution of traffic volumes based on the user throughput. In an embodiment of the present invention, the acquired statistical network counters and KPIs are aggregated data at the level of individual network cells. In particular, in the exemplary invention embodimentconsidered here, the acquired KPIs are or include the distribution of traffic volumesbased on the user throughput, and the average number of active users in the network cells. In particular, in an embodiment of the present invention, elementaryacquisition time intervals are defined, called “Records Of Period”, shortly “ROPs”,which are shorter than the first and second acquisition time intervals INTA and INTB,,lasting for example some minutes or tens of minutes each, e.g., 15 minutes. Each ofthe two acquisition time intervals INTA and INTB is made up of a set of successiveROPs: the first acquisition time interval INTA is made up of the succession of a numberP of ROPs ROPA,1, ROPA,2, …, ROPA,P; the second acquisition time interval INTB ismade up of the succession of a number Q of ROPs ROPB,1, ROPB,2, …, ROPB,Q. Thenumbers of ROPs in the two acquisition time intervals INTAand INTB(and their time duration) need not be the same. In every ROP ROPA,1, ROPA,2, …, ROPA,Pof the first acquisition time interval INTAand in every ROP ROPB,1, ROPB,2, …, ROPB,Qof the second acquisition time interval INTB the above-mentioned KPIs are acquired (distribution of traffic volumesper user throughput and average, in the ROP, of the number of active users in thenetwork cells).5. Activity 225In order to carry out a statistical characterization of the network traffic, a traffichomogeneity metric is defined so as to be then able to classify the ROPs during whichthe KPIs will be acquired in homogeneity classes, which are classes of homogeneityof the network traffic amount. The traffic homogeneity metric can be based onmeasurements of active users or traffic volumes; these pieces of information can thenbe aggregated at the level of sets of cells, particularly at the level of the cell cluster 130. An exemplary traffic homogeneity metric for the classification of the ROPs will be described later on.6. Activity 230The traffic homogeneity classes are defined based on the defined traffichomogeneity metric.7. Activity 235 – Block 710The KPIs are acquired in the ROPs ROPA,1, ROPA,2, …, ROPA,P of the firstacquisition time interval INTA. In particular, the acquired KPIs that are collected, during the generic ROP ROPA,1, ROPA,2, …, ROPA,Pof the first acquisition time interval INTA, at the levelof the individual network cells 115a and 115b of the cell cluster 130, are aggregatedto obtain corresponding KPIs aggregated at the level of the cell cluster 130: for example, in the generic ROP ROPA,1, ROPA,2, …, ROPA,Pof the first acquisition timeinterval INTA, the acquired KPIs related to the individual network cells 115a and 115bof the cell cluster 130 are summed up for all the network cells of the cell cluster.8. Activity 240The modified network configuration (135 in Fig. 1) determined by theexecution of the network optimization algorithm is implemented (140 in Fig. 1) in themobile network, modifying the cells parameters of one or more of the Target cells 115a.9. Activity 245 – Block 710Similarly to Activity 235, I KPIs are acquired in the ROPs ROPB,1, ROPB,2,…, ROPB,Q of the second acquisition time interval INTB after the networkconfiguration change. As before, the acquired KPIs collected, during the generic ROP ROPB,1,ROPB,2, …, ROPB,Q, at the level of the individual network cells 115a and 115b of thecell cluster 130, are aggregated (e.g., summed up) to obtain corresponding KPIs aggregated at the level of the cell cluster 130.10. Activity 250 – Blocks 715 and 720As schematized in Fig. 4, the generic ROPX,i (with X = A or B and i = 1, 2, …,P or i = 1, 2, …, Q) includes an aggregation, 405, at the level of the cell cluster 130and of the ROP, of KPIs collected, during the time interval of the ROP ROPX,i, at thelevel of the individual network cells 115a and 115b of the cell cluster 130. The KPIscollected at the level of the individual network cells of the cell cluster include the userthroughput ThrP and the number of active users Nu, averaged in the ROP. Theaggregation 405 results in an aggregated average number of users Nu[ROPX,i] in thecell cluster 130 in the ROP and in an aggregated user throughput distribution in thecell cluster 130 in the ROP, in terms of volume of carried traffic TrV (e.g., in kB) forthroughput value ThrP (e.g., in kbit / s). The aggregated average number of usersNu[ROPX,i] in the ROP is for example calculated as the sum, over the network cells ofthe cell cluster 130, of the average, in the ROP, of the number of active users. Fig. 5schematically depicts an exemplary aggregated user throughput distribution (merelyas an example, a gaussian distribution is shown). From the aggregated user throughputdistribution, the 5th percentile ThrP5%[ROPX,i] and the 50th percentile ThrP50%[ROPX,i](i.e., the median) of the aggregated user throughput distribution in the ROP are obtained. Every ROP (of the first and second acquisition time intervals INTA and INTB) is assigned to a respective one of the traffic homogeneity classes, as will be described in detail later on.11. Activity 255 – Block 725Based on the result of the classification of ROPs of the first and secondacquisition time intervals INTA and INTB according to traffic homogeneity classes,weights are calculated to be assigned to the ROPs of the second acquisition timeinterval INTB in the calculation of the network performance based on the network performance evaluation metric. The calculated weights are used to differently weight the ROPs of the second acquisition time interval INTBin the different traffic homogeneity classes.12. Activity 260 – Blocks 730 and 735Finally, the network performance evaluation metric is applied to the ROPs of the first acquisition time interval INTA, and to the weighted ROPs (ROPs weighted by the calculated weights) of the second acquisition time interval INTB, to obtain respective reward values. The two results of the network performance evaluationmetric, i.e., the two reward values are compared and it is thus assessed whether thenew network configuration identified by the optimization algorithm and implemented in the mobile network is effective in improving the mobile network performance, inparticular the QoS (output 740 in Fig. 7).As mentioned in the activities overview presented above, the method accordingto embodiments of the present invention envisages the definition of a metric for theevaluation of the network performance, particularly a metric useful for the assessment of the effectiveness of a network configuration change. In particular, the network performance evaluation metric is based on the QoS (QoS evaluation metric) and may take into account user throughput, number of handovers and success percentage,energy consumption. The QoS evaluation metric is based on a statistical aggregation,over an extended time interval (the first and second acquisition time intervals INTAand INTB, lasting some days or a few weeks, before and after the implementation ofthe network configuration change), of measurements, statistical network counters,KPIs acquired in the elementary acquisition time intervals, ROPs, which, in succession, make up the extended time interval. More specifically, in embodiments of the present invention the QoS evaluationmetric is applied to a set of network cells. In an embodiment of the present invention, the QoS evaluation metric can for example be defined by considering all the Targetcells 115a and Corona-1 cells 115b, i.e., the cell cluster 130, whose overall arearemains constant, in a first approximation, as the network configuration parameters(network cells parameters) vary as a consequence of the deployment of the modified(optimized) network configuration. In another embodiment of the present invention,the Corona-2 cells 115c can also be considered in the definition of the cell cluster andin the QoS evaluation metric. An exemplary, suitable metric for the evaluation of the QoS, when userthroughput is the quantity of importance, is a rewarding function (or rewarding metric)being a combination of two different percentiles of the user throughput distribution (see, e.g., Fig.5), where one of the two percentiles reflects the QoS (in terms of user throughput) experienced by the population of more penalized users and the other of the two percentiles reflects the QoS of a greater population of users. For example, the 5thpercentile of the user throughput distribution is considered, to reflect the QoS experienced by the population of more penalized users, and the 50thpercentile of theuser throughput distribution is considered, to reflect the QoS experienced by half ofthe population of users. For example, the rewarding function W can be expressed as:W = p1*ThrP5% + p2*ThrP50%which represents the weighted average between the average user throughput of theworst 5% of users (5th percentile of the user throughput distribution; but anotherpercentile values may be adopted here), ThrP5%, weighted by a first weight p1, and themedian user throughput of the 50% of users (50th percentile of the user throughputdistribution; another percentile value may be adopted here, too), Thr_ut50%, weightedby a second weight p2. The first and second weights p1 and p2 can for example bechosen to be equal to 0.5, although other values are possible (in any event, p1 + p2 =1). This rewarding function W is aimed at favouring the population of the more penalized users without too much negatively affecting the median conditions. The user throughput can be calculated in accordance to different approaches.According to an embodiment of the present invention, one method to calculatethe user throughput (in uplink, or in downlink, or overall, i.e., in uplink and indownlink) is the measurement, made by the mobile network, of the offered throughput,i.e., the throughput that the mobile network is able to provide to the users: the networkprovides, among the monitoring parameters at network cell level, the distributiondensity of user throughput (percentage and volumes of traffic per throughput valuesinterval); from these measurements it is possible to estimate the 5th percentile and the50th percentile of the user throughput distribution density. These percentiles are madeavailable for each individual ROP (of the first and second acquisition time intervals)at the level of individual cells 115 of the cell cluster 130 (Target cells 115a plusCorona-1 cells 115b).In order to evaluate the rewarding function W for each ROP (ROPs ROPA,1,ROPA,2, …, ROPA,P of the acquisition time interval INTA preceding the networkconfiguration change and ROPs ROPB,1, ROPB,2, …, ROPB,Q of the acquisition timeinterval INTB after the network configuration change), the distribution density of theuser throughput aggregated at the level of all the network cells of the cell cluster 130and in the considered ROP is considered, obtained by aggregating, for all the networkcells of the cell cluster 130, the distribution of traffic volumes as a function of thethroughput that the mobile network is able to provide to the user, thereby obtainingthe values Thr_ut5% and Thr_ut50% for each ROP (as pictorially shown in Fig. 4).According to another embodiment of the present invention, another method toestimate the user throughput is to start from estimates of SINR per pixel. The mobilenetwork capacity in the generic pixel is estimated, and the mobile network capacityestimation is then divided by the number of users included in the best server areaobtained from the measurements of each cell. The percentile values are evaluated afterattributing to each pixel a respective weight which is given by the intensity of thetraffic source in the pixel (e.g., the number of active users in the considered pixel).Compensations can be provided for users with worse SINR at the expense of thosewith higher SINR. The entire area of the cell cluster 130 is then considered and thevalues Thr_ut5% and Thr_ut50% are estimated.The number of users per cell plays an important role in evaluating the 5th and50th percentiles of the user throughput; in fact, a change in the offered traffic per cellwill substantially perturb these user throughput estimates. The estimate of the QoSthat the mobile network is able to provide to the users is in fact a function of the numberof active users distributed across the territory served (aggregated) in the individualcells: as the number of active users decreases, the 5th and 50th percentiles userthroughput increase, from which it follows that the network performance evaluation isvery sensitive to the distribution of the number of active users in the network cells and, therefore, it is sensitive to the distribution of the traffic volumes in the area of the cellcluster 130. This makes it very difficult to compare acquisitions of throughputmeasurements carried out in different time intervals.The QoS undergoes degradation in the presence of non-homogeneity in the number of active users from network cell to network cell, and the rewarding functionW is capable of detecting this. With greater dissymmetry, the 5th percentile throughputand the 50th percentile throughput will tend to degrade, to the advantage of networkcells with a low number of users with the higher percentiles.The coherence between traffic volumes (number of users) at the cell level isnormally not maintained with the change of network configuration: the acquisition ofnetwork measurements in two different time periods relating to the time before and thetime after the network configuration change will be affected by the variation in thenetwork cells’ coverage area (resulting from the change of the network configuration)and therefore the variation in the number of active users per network cell, even withthe same offered traffic. However, the Applicant observed that it is possible to at leastguarantee consistency of traffic volumes at the level of coverage area of the cell cluster130, as the cluster coverage area (overall coverage area of the Target cells 115a andthe Corona-1 cells 115b) remains, in first approximation, constant as the network cells’parameters change. It is then possible to impose a constraint on the total traffic at the cell cluster level in the generic ROP in order to compare consistent ROPs with respectto the traffic volume / traffic amount.According to an embodiment of the present invention, a comparison on thenetwork measurements, network statistical counters, KPIs, particularly KPIs related tothe user throughput distribution, is carried out by considering time periods withhomogeneous volumes / amounts of traffic at level of the cell cluster 130.Considering for example network cells’ measurements with ROPs of 15minutes, the constraint about homogeneous traffic volumes / amounts translates into amodification of the rewarding function W applied in the processing of the acquiredKPIs (particularly, the user throughput distribution) in ROPs ROPB,1, ROPB,2, …, ROPB,Qof the acquisition time interval INTBafter the network configuration change, so as to compare ROPs ROPA,1, ROPA,2, …, ROPA,Qof the acquisition time interval INTA and ROPs ROPB,1, ROPB,2, …, ROPB,Q of the acquisition time interval INTBwhich feature homogeneous traffic sources in the cell cluster 130.According to an embodiment of the present invention, e.g., KPIs relating to thenetwork cells of the cell cluster 130 are acquired in the unitary time intervals ROPsROPA,1, ROPA,2, …, ROPA,P of the acquisition time interval INTA before the time tc of the network configuration change and in the ROPs ROPB,1, ROPB,2, …, ROPB,Qof the acquisition time interval INTB after the network configuration change. In an embodiment of the present invention, the acquired KPIs includemeasurements of the traffic volumes as a function of the user throughput that themobile network is able to provide to the users (user throughput distribution), andmeasurements of the number of active users (averaged) in the ROPs. These KPIs canbe acquired at the level of the individual network cells of the cell cluster 130, and maythen be aggregated to obtain values relating to the cell cluster 130 as a whole.As previously mentioned, the first and second acquisition time intervals INTA,and INTB may last some days or a few weeks and may include different numbers ofROPs. Let P be the number of ROPs making up the acquisition time intervals INTAbefore the network configuration change and Q be the number of ROPs ROPB,1,ROPB,2, …, ROPB,Q making up the acquisition time interval INTB after the networkconfiguration change, with P ^ Q.For each ROP ROPX,i (with X = A or B and i = 1, 2, …, P or i = 1, 2, …, Q),the overall number of active users Nu in the cell cluster 130 and in the ROP isevaluated, and the 5th and 50th percentiles of the user throughput distribution,aggregated on the cell cluster 130, are calculated.According to an embodiment of the present invention, traffic homogeneityclasses are defined, and every ROP ROPX,i is then assigned to a respective one of thedefined traffic homogeneity classes. According to an embodiment of the present invention, the definition of traffichomogeneity classes includes considering as an indication of the traffic volume / amount the number of active users in the cell cluster 130, and identifying the minimumof the overall number of active users Nu,min in the cell cluster 130 and the maximum ofthe overall number of active users Nu,max in the cell cluster 130 in the first acquisitiontime interval INTA. The range of values Nu,max] is subdivided into apredetermined number n of sub-ranges or bins. For example, the number n of bins canbe 20, with a variability of the number of active users (i.e., of the traffic volume) ofless than approximately 5% within each bin. In another example, with reference toFig.6A – Fig.6C, the range of values [Nu,min, Nu,max] is subdivided into five bins BinA,1– BinA,5, from the minimum of the overall number of active users Nu,min measured, inthe first acquisition time interval INTA, in the cell cluster 130 to the maximum of theoverall number of active users Nu,maxmeasured, in the acquisition time interval INTAbefore the network configuration change, in the cell cluster 130. Every ROP ROPA,i (i = 1, 2, …, P) of the first acquisition time interval INTAis assigned to a respective one of the traffic homogeneity classes defined as above, i.e.,to one of the bins BinA,1 – BinA,5, based on the value of the parameter Nu[ROPA,i] forthat ROP. Fig. 6A pictorially shows the classification of the ROPs ROPA,i (i = 1, 2, …,P) of the first acquisition time interval INTA: in the example, three ROPs are classifiedin bin BinA,1, four ROPs are classified in bin BinA,2, six ROPs are classified in bin BinA,3, five ROPs are classified in bin BinA,4, and two ROPs are classified in bin BinA,5(in a practical scenario, the number of ROPs making up the first acquisition time interval INTAwill be higher: with ROPs of 15 minutes each, one day includes 96ROPs, so that a first acquisition time interval INTA of one week includes 672 ROPs;possibly, also the number of bins – number of traffic homogeneity classes will begreater than 5). The piecewise linear curve shown in dotted line and identified as 605Ain Fig. 6A represents the profile of traffic distribution in the cell cluster 130, in theacquisition time interval INTA. Similarly, every ROP ROPB,j (j = 1, 2, …, Q) of the second acquisition timeinterval INTB is assigned to a respective one of the traffic homogeneity classes definedas above, i.e., to one of the bins BinB,1 – BinB,5, (which correspond, in terms of rangesof values of number of active users in the interval Nu,max], to the bins BinA,1–BinA,5) based on the value of the parameter Nu[ROPB,j] for that ROP.Fig. 6B pictorially show the classification of the ROPs ROPB,j (j = 1, 2, …, Q)of the second acquisition time interval INTB: in the example, two ROPs are classified in bin BinB,1, one ROP is classified in bin BinB,2, four ROPs are classified in bin BinB,3, six ROPs are classified in bin BinB,4, and five ROPs are classified in bin BinB,5. The piecewise linear curve shown in dotted line and identified as 605B in Fig. 6B represents the profile of traffic distribution in the cell cluster 130, in the second acquisition time interval INTB. In order to assess the effectiveness, or “goodness”, of the modified network configuration that has been identified by the network optimization algorithm and that, at the time tc, has been implemented in the mobile network (by modifying the cellparameters of the one or more of the Target cells 115a of the cell cluster 130), a rewardis calculated and assigned to the network performance, i.e., to the QoS (in terms ofuser throughput) in respect of the network configuration before the networkconfiguration change. This can for example be done by calculating the value WA of arewarding function, for example with the formula of the rewarding function Wpresented in the foregoing, on the ROPs ROPA,i (i = 1, 2, …, P) of the first acquisitiontime interval INTA, particularly on the values ThrP5% (5thpercentile of the throughput distribution) and ThrP50% (5thpercentile of the throughput distribution) of the ROPsROPA,i (i = 1, 2, …, P):WA = p1*ThrP5% + p2*ThrP50% One possibility to calculate the rewarding function WA for the network configuration before the network configuration change is to perform the calculation of the reward WA,ifor every ROP ROPA,iindividually, and then calculating the average of the calculated rewards reward WA,i. Another way to calculate the rewarding functionW for the network configuration before the network configuration change is to obtainan aggregated user throughput distribution, aggregated over all the ROPs ROPA,i, then calculate the 5thand 50thpercentiles of the aggregated user throughput distribution and apply to calculated percentiles the formula WA = p1*ThrP5% + p2*ThrP50%. Concerning the network performance, i.e., the QoS (in terms of userthroughput) after the change in the network configuration, it can be seen that the trafficdistribution profile 605B in the cell cluster 130 in the second acquisition time intervalINTB is different from the traffic distribution profile 605A in the cell cluster 130 in thefirst acquisition time interval INTA, this being the consequence of random factors likethe mobility in time of the users through the coverage area of the cell cluster 130 (i.e.,of the traffic sources). The Applicant understood that such differences in the trafficdistribution profiles before and after the network configuration change render a directcomparison of network measurements (KPIs) before and after the networkconfiguration change scarcely reliable in the assessment of the effectiveness, or ineffectiveness, of the network optimization action identified by the network optimization algorithm and implemented on field. In other words, calculating therewarding function W directly on the measurements / KPIs acquired in the ROPsROPB,j (j = 1, 2, …, Q) of the second acquisition time interval INTB without takinginto consideration the variations in the traffic distribution would provide a value that is scarcely significant for a comparison with the value WA calculated for the ROPsROPA,i (i = 1, 2, …, P) of the first acquisition time interval INTA.According to the present invention, for the calculation of the reward to be assigned to the modified network configuration which has been identified by the network optimization algorithm and implemented in the mobile network, the trafficdistribution profile, calculated in respect of the measurements acquired during thesecond acquisition time interval INTB, exemplified by the curve 605Bin Fig. 6B, is “re-shaped” to correspond, i.e., to be rendered similar to the traffic distribution profilecalculated in respect of the measurements acquired during the first acquisition timeinterval INTA, exemplified by the curve 605Ain Fig.6A. According to an embodiment of the present invention, the re-shaping of the traffic distribution profile related to the second acquisition time interval INTB, curve 605B in Fig. 6B, is based on a statistical processing on the network measurements inthe ROPs ROPB,j (j = 1, 2, …, Q) of the second acquisition time interval INTB.According to an embodiment of the present invention, the statistical processingon the network measurements in the ROPs ROPB,j (j = 1, 2, …, Q) of the secondacquisition time interval INTBis for example performed as described herebelow. The aggregated measurements in the ROPs ROPB,j (j = 1, 2, …, Q) of theacquisition time interval INTB after the network configuration change, particularly thevalues of the 5th percentile ThrP5%[ROPX,i] and of the 50th percentile ThrP50%[ROPX,i]of the aggregated user throughput distribution in the ROPs ROPB,j (j = 1, 2, …, Q) areweighted by respective weights. According to an embodiment of the present invention, the weights WeightB,j tobe assigned to the ROPs ROPB,j (j = 1, 2, …, Q) of the acquisition time interval INTBfor the re-shaping of the traffic distribution profile related to the second acquisition time interval INTB and then in the computation of the rewarding function WB are determined as: WeightB,j [BinB,k] = #BinA,k / (#BinB,k * P) where: -WeightB,j [BinB,k] is the weight to be assigned to the generic ROP ROPB,j (j =1, 2, …, Q) of the second acquisition time interval INTBwhich has been classified in the bin BinB,k (with k = 1, …, MaxBin, MaxBin = 5 in the exampleof Fig. 6B, that identifies one of the bins BinB,1 – BinB,5, i.e., one of the traffichomogeneity classes defined above); -#BinA,k (with k = 1, …, MaxBin, MaxBin = 5 in the example of Fig. 6A, thatidentifies one of the bins BinA,1 – BinA,5, i.e., one of the traffic homogeneityclasses defined above) is the cardinality of the bin BinA,kin the classificationof the ROPs ROPA,i (i = 1, 2, …, P) of the first acquisition time interval INTA,i.e., the number of ROPs ROPA,i (i = 1, 2, …, Q) that have been classified inthe bin BinA,k; -#BinB,k is the cardinality of the bin BinB,k, and- P is the overall number of ROPs in the first acquisition time interval INTA.Back to the example of Fig. 6A and Fig. 6B, the overall number P of ROPs inthe first acquisition time interval INTA is 20, P = 20, and the weights are:#BinA,1 = 3 #BinB,1 = 2 WeightB,i [BinB,1] = 3 / (2*20) = 0.075#BinA,2 = 4 #BinB,2 = 1 WeightB,i [BinB,2] = 4 / (1*20) = 0.2#BinA,3 = 6 #BinB,3 = 4 WeightB,i [BinB,3] = 6 / (4*20) = 0.075#BinA,4 = 5 #BinB,4 = 6 WeightB,i [BinB,4] = 5 / (6*20) = 0,042#BinA,5 = 2 #BinB,5 = 5 WeightB,i [BinB,5] = 2 / (5*20) = 0,02In case no ROPs ROPB,j (j = 1, 2, …, Q) of the second acquisition time intervalINTB have been classified in one or more of the bins BinB,1 - BinB,5, said bin or binsare not taken into consideration, i.e., the number of ROPs of the first acquisition timeinterval INTA which have been classified in those bin or bins BinA,1 – BinA,5corresponding to the empty bin or bins BinB,1 - BinB,5 is subtracted from the overallnumber of ROPs P.The value of the rewarding function WBis then calculated based on the values [ThrP5%]ϯand [ThrP50%]ϯ, which are the values ThrP5% (5thpercentile of the throughput distribution) and ThrP50%(5thpercentile of the throughput distribution) ofthe ROPs ROPB,j (j = 1, 2, …, Q), weighted by the weights WeightB,j [BinB,k]:WB = p1*[ThrP5%]ϯ+ p2*[ThrP50%]ϯHere again there are two possibilities for calculating the rewarding function WBfor the network configuration after the network configuration change: one possibilityis to perform the calculation of the reward WB,j for every ROP ROPB,j (weighted by the respective weight WeightB,j [BinB,k]) individually, and then calculating the average of the calculated rewards reward WB,j; another possibility is to obtain an aggregated user throughput distribution, aggregated over all the ROPs ROPB,j(weighted by therespective weight WeightB,j [BinB,k]) and then calculate the 5th and 50th percentiles ofthe aggregated (weighted) user throughput distribution and apply to calculated percentiles the formula WB = p1*[ThrP5%]ϯ+ p2*[ThrP50%]ϯ. In particular, theaggregated user throughput distribution can be calculated by weighting the userthroughput distribution of each ROP ROPB,j by the respective weight WeightB,j [BinB,k] = #BinA,k / (#BinB,k * P). The calculated reward function value WBis then compared to the calculated reward function value WA, to assess whether the modified network configuration implemented on the field has brought about benefits, in terms of QoS, compared to the previous network configuration. For example, a value WBsufficiently higher than the value WA ensures that the modified network configuration has improved the QoS. In this way, a homogenous statistical characterization is given to the acquired network measurements, statistical network counters, KPIs, e.g., of user throughput,acquired at different times (before and after the network configuration change), so thatthe acquired network measurements can be meaningfully compared with each other. The homogeneous statistical characterization can be given based on measurements of the number of active users in the cell cluster 130, attributing suitable weights to the individual ROPs, suitable to guarantee such homogeneity. The weights attributed to the ROPs are aimed at enabling a comparison between ROPs with homogeneous numbers of active users in the cell cluster. Alternative formulations of the rewarding function W are possible.If, based on the comparison of the reward function values WA and WB, it results that the modified network configuration implemented on the field has not brought about benefits, in terms of QoS, compared to the previous network configuration (because the QoS after the change has not significantly improved, not to say the QoS after the change is worse than before), then, one or more of the following actions may be undertaken. One possible action is to re-run the optimization / simulation algorithm by modifying: -the ranges of variability of the modifiable parameters (e.g., antennas’electrical tilts and / or azimuth, transmitted power, network cells’ selection / re-selection thresholds), so as to increase the degrees of freedom of the networkoptimization / simulation algorithm;- the set of modifiable parameters that the network optimization / simulationalgorithm can consider (if for example in the previous run of theoptimization / simulation algorithm not all the possible modifiable parameters have been taken into consideration); -the set of network cells (Target cells 115a) whose parameters can bemodified; -the importance (weights p1, p2) assigned to the different items (e.g., userthroughputs ThrP5%, ThrP50%) considered in the calculation of the reward function for the network configurations before and after the change, thereby modifying the goals set to the optimization / simulation algorithm; -as a further resource, modifying the optimization / simulation algorithm (forexample, in case of optimization / simulation algorithms based on machine learning / artificial intelligence, this may involve a re-training of the algorithm). In accordance with another embodiment of the present invention, in addition to the first and second acquisition time intervals INTA and INTB described in theforegoing, a further acquisition time interval INTR is defined, referred to as referenceacquisition time interval INTR, during which network measurements and KPIs are acquired. The reference acquisition time interval INTRis a time interval selected sufficiently long (the duration of some days or a few weeks) so as to represent a genericcondition of the mobile network (which is taken as a reference mobile networkcondition), i.e., a condition not biased by particular events that may occur in peculiarmoments of time, before and / or after the network configuration change. As depictedin Fig. 8, the reference acquisition time interval INTR may for example be a timeinterval starting before (moment tR1) the moment tc of the implementation of thenetwork configuration change and ending after (moment tR2) the moment tc of the network configuration change; for example (but this is not limiting), the reference acquisition time interval INTRembraces the first and second acquisition time intervalsINTA and INTB (in the example of Fig. 8 the start (moment tR1) of the referenceacquisition time interval INTR is before the start (moment tA1) of the first acquisition time interval INTA and the end (moment tR2) of the reference acquisition time interval INTRis after the end (moment tB2) of the second acquisition time interval INTB; however, the reference acquisition time interval INTR might as well start at the beginning of the first acquisition time interval INTA and terminate at the end of the second acquisition time interval INTB. A proper selection of the reference acquisitiontime interval INTR, e.g. a relatively long time interval, ensures that it reliably representa generic condition of the mobile network, i.e., a robust reference mobile network condition. Like the two acquisition time intervals INTA and INTB, the reference acquisition time interval INTR is made up of a succession of reference ROPs ROPR,1, ROPR,2, …, ROPR,N, for example lasting 15 minutes each. The network measurements, statistical network counters and KPIs, particularlythe network traffic / number of active users and the user throughput distribution, acquired during the reference acquisition time interval INTR are exploited to obtain a reference statistical characterization of the network traffic (i.e., differently from the previously described invention embodiment, the statistical characterization of the network traffic is not derived from the KPIs acquired during the first acquisition time interval INTA). The traffic homogeneity classes for the statistical characterization of thenetwork traffic are defined in respect of the reference acquisition time interval INTR.As in the previously described invention embodiment, the number of active users inthe cell cluster 130 during the reference acquisition time interval INTR is consideredas an indication of the traffic volume, and the minimum of the overall number of activeusers NuR,min in the cell cluster 130 and the maximum of the overall number of activeusers NuR,max in the cell cluster 130 in the reference acquisition time interval INTR areidentified. The range of values [NuR,min, NuR,max] is subdivided into a predeterminednumber n of sub-ranges or bins, for example 20, with a variability of the number ofactive users (i.e., of the traffic) of less than approximately 5% within each bin. In adifferent example, with reference to Fig. 9A, the range of values [NuR,min, NuR,max] issubdivided into five bins BinR,1 – BinR,5, from the minimum of the overall number ofactive users NuR,min measured, in the reference acquisition time interval INTR, in thecell cluster 130 to the maximum of the overall number of active users NuR,max measured,in the reference acquisition time interval INTR, in the cell cluster 130. Every reference ROP ROPR,h (h = 1, 2, …, N) of the reference acquisition timeinterval INTR is assigned to a respective one of the traffic homogeneity classes definedas above, i.e., to one of the bins BinR,1 – BinR,5, based on the value of the parameterNu[ROPR,h] (average number of active users in the reference ROP) for that referenceROP. Fig. 9A pictorially shows an exemplary classification of the reference ROPsROPR,h (h = 1, 2, …, N) of the reference acquisition time interval INTR. In theexample, three reference ROPs are classified in bin BinR,1, four reference ROPs are classified in bin BinR,2, six reference ROPs are classified in bin BinR,3, five reference ROPs are classified in bin BinR,4, and two reference ROPs are classified in bin BinR,5 (as mentioned in connection with the previously described invention embodiment, ina practical scenario the number of reference ROPs ROPR,h, h = 1, 2, …, N, making upthe reference acquisition time interval INTR will be higher: with reference ROPs of 15 minutes each, one day includes 96 reference ROPs, so that a reference acquisition time interval INTR of one week includes 672 reference ROPs). The piecewise linear curveshown in dotted line and identified as 905R in Fig. 9A represents the profile ofreference traffic distribution in the cell cluster 130, in the reference acquisition time interval INTR. Similarly, every ROP ROPA,i (i = 1, 2, …, P) of the first acquisition timeinterval INTA, and every ROP ROPB,j (j = 1, 2, …, Q) of the second acquisition timeinterval INTB is assigned to a respective one of the traffic homogeneity classes definedas above, i.e., to one of the bins BinR,1 – BinR,5 (which correspond to ranges of valuesof the number of active users in the interval [NuR,min, NuR,max]), based on the value ofthe parameter Nu[ROPX,i] (X = A or B, i = 1, …, P or 1, …, Q) for that ROP.Fig. 9B and Fig. 9C pictorially show the classification of the ROPs ROPA,i (i= 1, 2, …, P) of the acquisition time interval INTA before the network configurationchange and the classification of the ROPs ROPB,j (j = 1, 2, …, Q) of the acquisitiontime interval INTB after the network configuration change exploiting the traffichomogeneity classes BinR,1 – BinR,5 defined on the basis of the reference acquisitiontime interval INTR: in the example: -for the acquisition time interval INTA before the network configurationchange four ROPs are classified in bin BinA,1(corresponding to bin BinR,1), two ROPs are classified in bin BinA,2 (corresponding to bin BinR,2), three ROPs are classified inbin BinA,3 (corresponding to bin BinR,3), five ROPs are classified in bin BinA,4(corresponding to bin BinR,4), and four ROPs are classified in bin BinA,5(corresponding to bin BinR,5). The piecewise linear curve shown in dotted line andidentified as 905A in Fig. 9B represents the profile of traffic distribution in the cellcluster 130, in the acquisition time interval INTA; -for the acquisition time interval INTB after the network configuration changetwo ROPs are classified in bin BinB,1 (corresponding to bin BinR,1), one ROP is classified in bin BinB,2 (corresponding to bin BinR,2), four ROPs are classified in bin BinR,3(corresponding to bin BinR,3), six ROPs are classified in bin BinB,4(corresponding to bin BinR,4), and five ROPs are classified in bin BinB,5 (corresponding to bin BinR,5). The piecewise linear curve shown in dotted line andidentified as 905B in Fig. 9C represents the profile of traffic distribution in the cellcluster 130, in the acquisition time interval INTB. It can be seen that the traffic distribution profiles 905A and 905B in the cellcluster 130 in the acquisition time intervals INTA and INTB are different from thereference traffic distribution profile 905R in the cell cluster 130 in the referenceacquisition time interval INTR, this being the consequence of random factors like themobility in time of the users through the coverage area of the cell cluster 130 (i.e., ofthe traffic sources). As in the previously described invention embodiment, according to the present invention, for the calculation of the reward to be assigned to the modified network configuration which has been identified by the network optimization algorithm and implemented in the mobile network, the traffic distribution profiles 905Aand 905B, calculated on the measurements acquired during the first acquisition time interval INTAand, respectively, during the second acquisition time interval INTB, are “re- shaped” to be rendered similar to the reference traffic distribution profile 905Rcalculated on the measurements acquired during the reference acquisition time interval INTR. According to an embodiment of the present invention, the re-shaping of thetraffic distribution profiles 905A and 905B related to the first and second acquisitiontime intervals INTA and INTB is based on a statistical processing on the networkmeasurements, statistical network counters, KPIs collected in the ROPs ROPA,i (i = 1,2, …, P) of the first acquisition time interval INTA and in the ROPs ROPB,j (j = 1, 2,…, Q) of the second acquisition time interval INTB.According to an embodiment of the present invention, the statistical processingon the KPIs in the ROPs ROPA,i (i = 1, 2, …, P) and ROPB,j (j = 1, 2, …, Q) is forexample performed as described herebelow. The aggregated KPIs in the ROPs ROPA,i (i = 1, 2, …, P) and ROPB,j (j = 1,2, …, Q), particularly the values of the 5th percentile ThrP5%[ROPY,i] and of the 50thpercentile ThrP50%[ROPY,i] of the aggregated user throughput distribution in the ROPsare weighted by respective weights. The weights WeightY,w (Y = A or B; w = 1, …, P for Y = A, w = 1, …, Q for Y =B) to be assigned to the ROPs ROPA,i (i = 1, 2, …, P) and ROPB,j (j = 1, 2, …, Q) inthe computation of the rewarding functions WA and WB are determined as: WeightY,w[BinY,k] = #BinR,k / (#BinY,k* N) where: -WeightY,w [BinY,k] is the weight to be assigned to the generic ROP ROPA,i (i =1, 2, …, P) or ROPB,j (j = 1, 2, …, Q) which has been classified in the binBinY,k (with k = 1, …, MaxBin, MaxBin = 5 in the example of Fig. 9A, thatidentifies one of the bins BinY,1 – BinY,5, i.e., one of the traffic homogeneityclasses defined above); -#BinY,k (with k = 1, …, MaxBin, MaxBin = 5 in the example of Fig. 9A, thatidentifies one of the bins BinY,1 – BinY,5, i.e., one of the traffic homogeneityclasses defined above) is the cardinality of the bin BinY,kin the classification of the ROPs ROPA,i (i = 1, 2, …, P) and ROPB,j (j = 1, 2, …, Q), i.e., thenumber of ROPs ROPA,i (i = 1, 2, …, P) and ROPs ROPB,j (j = 1, 2, …, Q)that have been classified in the bin BinY,k; -#BinR,k is the cardinality of the bin BinR,k, and- N is the overall number of reference ROPs in the reference acquisition timeinterval INTR. Similarly to the previously described embodiment, in case no ROPs ROPA,i(i= 1, 2, …, P) or ROPs ROPB,j (j = 1, 2, …, Q) have been classified in one or more ofthe bins BinA,1 – BinA,5 or BinB,1 - BinB,5, said bin or bins are not taken intoconsideration, , i.e., the number of ROPs of the reference acquisition time interval INTR which have been classified in those bin or bins corresponding to the empty binor bins BinA,1 – BinA,5 or BinB,1 - BinB,5 is subtracted from the overall number of ROPsN. The values of the rewarding functions WA and WB are then calculated based onthe values [ThrP5%]Aϯand [ThrP50%]Aϯ, [ThrP5%]Bϯand [ThrP50%]Bϯ, which are the values ThrP5% (5thpercentile of the throughput distribution) and ThrP50% (50thpercentile of the throughput distribution) of the ROPs ROPA,i (i = 1, 2, …, P) andROPB,j (j = 1, 2, …, Q), weighted by the weights WeightY,w [BinY,k]:WA = p1*[ThrP5%]Aϯ+ p2*[ThrP50%]AϯWB= p1*[ThrP5%]Bϯ+ p2*[ThrP50%]BϯAs before, any of the ways described for calculating the reward WA and WB can be adopted. The calculated reward WBfor the network configuration after the network configuration change is than compared to the calculated reward WA for the network configuration before the network configuration change, to assess whether the modified network configuration implemented on the field has brought about benefits, in terms of QoS, compared to the previous network configuration. For example, a value WB sufficiently higher than the value WAensures that the modified network configuration has improved the QoS. Essentially, a difference of this embodiment with respect to the previously described embodiment is that in the previously described embodiment the first acquisition time interval INTAis taken as the reference acquisition time interval, andthe weights WeightA,w (w = 1, …, P) are unitary weights.^ ^ ^ In the invention embodiment described in the foregoing, the overall traffic (forexample, expressed in terms of number of users, or average number of active users) inthe cell cluster 130 as a whole has been considered in the definition of classes of trafficvolume. The acquired KPIs to be used for homogenizing the network traffic(distribution of traffic sources, i.e., users) before and after the network configurationchange are data (number of active users, user throughput distribution) that the mobile network provides aggregated at the level of the network cells of the cell cluster 130 and which are then aggregated at the level of the cell cluster 130, and the classes of network traffic are classes of network traffic at the level of the cell cluster 130. An alternative method for network traffic homogenization will be nowdescribed, which is based on the geographic distribution of the volumes of traffic (e.g., number of active users) across the territory covered by the cell cluster 130. As before, in order to assess the benefits, in terms of QoS (and, particularly, QoS in terms of user throughput) deriving from the change of the network configuration, the user throughput before and after the network configuration change to be compared should be referred to homogeneous conditions of network traffic volume (e.g., expressed in terms of average number of active users), so that the measured user throughput before and after is actually comparable. While in the previously described embodiment the overall network traffic in the cell cluster 130 was considered, in this alternative method a finer geographic distribution of the network traffic volume is considered, at the level of pixels. As described in detail below, in this alternative network traffic homogenizationmethod, classes of homogeneous traffic amounts are defined based on distribution ofusers in portions, sub-areas of the network cells’ areas, e.g., in pixels of the geographic territory, exploiting georeferenced measurements. Indicators of network traffic amount may be any one of, or one or more of: indicators of network traffic volume (e.g., expressed in kbits), indicators of userpresences in the network cells (i.e., users for which the mobile network records a lastevent in the network cells, derived from indications about the location of the user whenthe last event has been recorded), indicators of numbers of active users in the networkcells. In particular, this alternative network traffic homogenization method providesan alternative implementation of Activity 225 and Activity 230 described in theforegoing. Firstly, using for example a simulation tool (e.g., an electromagnetic fieldpropagation simulator, of the type often used by Telco during the planning phase of amobile communications network, capable of simulating the propagation of radiosignals through a territory taking into account data describing the territory, like naturalorography, presence of human artefacts, i.e., buildings and so on, presence of treesetc.,), or from geo-referenced measurements from the field, theoretical, hypotheticalcoverage areas of the network cells of the cell cluster 130 (i.e., the Target cells 115aplus the Corona-1 cells 115b) are evaluated, for the network configuration before thenetwork configuration change (first network configuration A) and for the new networkconfiguration after the network configuration change (second network configurationB). For each network cell of the cell cluster 130, three areas or zones Z1, Z2 andZ3 of the cell coverage area are defined as follows (as schematized in Fig. 10):- a first zone Z1: includes those pixels of the cell coverage area which fallwithin the cell’s best server area both in the first network configuration A and in secondnetwork configuration B, i.e., both before and after the network configuration change; -a second zone Z2: includes those pixels that fall within the cell’s best serverarea only in the first network configuration A (i.e., only before the network configuration change: these are lost pixels, i.e., pixels that are lost by the network cellafter the network configuration is changed from the first network configuration A tothe second network configuration B); -a third zone Z3: includes those pixels that fall within the cell’s best server areaonly in the second network configuration B (i.e., only after the network configurationchange: these are acquired pixels, i.e., pixels that are acquired by the network cell afterthe network configuration is changed from the first network configuration A to thesecond network configuration B). For each network cell of the cell cluster 130, during the ROPs of the first andsecond acquisition time intervals INTAand INTB, respectively before and after the implementation of the modified network configuration (and, where defined, also for the reference acquisition time interval INTRdescribed in the foregoing), geo- referenced measurements of network traffic volumes in the network cell are obtained (the measurements of network traffic volumes may be or include indications of network traffic volumes, such as measurements of user presences in the network cell– i.e., those users for which the mobile network records a last event in the network cell- and / or number of active users in the network cell).The measurements of network traffic volumes can be derived from MDTmeasurement campaigns (or other geo-referenced measurements). As known to thoseskilled in the art, MDT is a 3GPP standardized mechanism designed to Telcos toexploit user equipment of users in a mobile network to collect mobile network data:user equipment collect field measurements, including radio measurements; the MDT measurements are linked with information which makes it possible to derive the userequipment location, and a time stamp, at the time the field measurements are collected.The measurements of network traffic volumes can also be derived byaggregation of network statistical counters, e.g., counters per beam (5G mobilenetworks, having base stations equipped with smart antennas – i.e., highly directiveantennas – make available individual counters for each beam of directional antennas,providing a geographic detail in the network cells). The measurements are acquired during the ROPs ROPA,I (i = 1, 2, …, P) of the first acquisition time interval INTA, before the implementation of the modifiednetwork configuration, and the ROPs ROPB,i (i = 1, 2, …, Q) of the second acquisitiontime interval INTB, after the implementation of the modified network configuration (and, where defined, also for the reference acquisition time interval INTR). For each network cell of the cell cluster 130, the area of the union zone Z2^Z3being the union of the respective second zone Z2 and third zone Z3 is then considered,and the network traffic volume / user presences / number of active users in the network cell is calculated as the difference between the network traffic volume / user presences / number of active users in the second zone Z2 and the network traffic volume / userpresences / number of active users in the third zone Z3: the network traffic volume / user presences / number of active users thus calculated represent the network traffic volume / user presences / number of active users that the network cell has given away,lost (if the difference is a negative number) or acquired (if the difference is a positivenumber). In this way, for each network cell of the cell cluster 130 there is provided aquantification of the variation of the traffic volume before and after the network configuration change. This will enable to perform a comparison of the measurements before and after the network configuration change in areas of the network cells in which the network traffic remains substantially constant before and after the change in network configuration. Then, for each network cell, ranges of values (bins) of the network traffic volume / user presences / number of active users measured in the first acquisition timeinterval INTA (before the network configuration change) are defined for classificationpurposes, for the first zone Z1 and for the union zone Z2^Z3. For example, two binsfor the first zone Z1 and two bins for the union zone Z2^Z3 are defined: for example,one of the two bins may correspond to “higher” values of network traffic volume / userpresences / number of active users, the other bin may correspond to “lower” values of network traffic volume / user presences / number of active users, where “higher” and “lower” may be referred to a selected threshold of network traffic volume / userpresences / number of active users). Fig. 11 schematizes the definition of two bins,BinL and BinH, for the generic zone Z1 and the generic union zone Z2^Z3 of thegeneric network cell of the cell cluster 130; the bins are defined taking into consideration the number of active users Nu; Nu,Th denotes a threshold for the number of active users used to discriminate “higher” numbers of active users from “lower” numbers of active users. In principle, nothing prevents from defining more than two bins for theclassification of the measured network traffic volume / user presences / number of active users, however since in this invention embodiment the classification is made at the level of zones of every network cell (instead of at the cell cluster 130 level), it maybe preferable to keep the number of bins relatively low, for the sake of lower computational burden. The bins are used to define homogeneity classes of network traffic volume / user presences / number of active users. For the zones Z1 and the union zones Z2^Z3,the ROPs of the first acquisition time interval INTA and the ROPs of the second acquisition time interval INTB (and, where defined, also the ROPs ROPR,k, k = 1, 2,…, N, of the reference acquisition time interval INTR) will be classified in the properbin BinL or BinH, depending on, for example, the average number of active users in the ROPs. Considering all the Target cells 115a and, optionally, for the Corona-1 cells115b (or at least for a number v of more significant Corona-1 cells 115b which aremore significant in terms of traffic volumes / user presences / number of active users in the respective zone Z1 and / or union zone Z2^Z3: for example, Corona-1 cells 115bwith a traffic volume higher than 3 Erlang, or with a number of active users higher than a selected number of active users are considered), all the possible combinationsof the two bins BinL, BinH of the respective zone Z1 and union zones Z2^Z3 areidentified. The overall number of such combinations is: where #bins is the number of bins (two in the considered example, BinL or BinH), 2 isthe number of zones considered for each network cell (the first zone Z1 and the unionzone Z2^Z3), #Targetcells is the number of Target cells 115a in the cell cluster 130,and v[Corona-1cells] is the number v of more significant Corona-1 cells 115b of thecell cluster 130. The overall number of possible combinations of bins of the considered networkcells of the cell cluster 130 may easily become relatively high. In order to limit theoverall number of possible combinations of bins, the number of network cells may be reduced, for example by merging network cells, particularly by merging first zones Z1and / or union zones Z2^Z3 of different network cells, for example based on thenetwork traffic volume (merging together network cells having high traffic volumes, and cells having low traffic volumes). In particular, the average peak network trafficin the peak network traffic time may be considered: the first zones Z1 of the networkcells having high peak network traffic are merged together, and the first zones Z1 ofthe network cells having low peak network traffic are merged together; the union zoneZ2^Z3 of each network cell is partitioned into sub-areas which give away pixels tonetwork cells with high peak network traffic, and sub-areas which acquire pixels from network cells with low peak network traffic, then the sub-areas are aggregated based on the peak traffic volume. In connection with the reduction of the overall number of possiblecombinations of bins, it is observed that the union zones Z2^Z3 of the network cellscan be partitioned into sub-zones by considering the traffic volume (number of users) of the network cells that acquire / lose network traffic. Considering a network cell, theunion zone Z2^Z3 can be partitioned in sub-zones, each of which represents anaggregate of pixels that “migrate” from (second zone Z2) and to (third zone Z3) the considered network cell, respectively to and from network cells with classes ofnetwork traffic which are homogenous (i.e., pixels of the second zone Z2 of theconsidered network cell that migrate to other network cells having low traffic volume,and pixels of the third zone Z3 of the considered network cell that come from othernetwork cells having low traffic volume; and pixels of the second zone Z2 of theconsidered network cell that migrate to other network cells having high traffic volume,and pixels of the third zone Z3 of the considered network cell that come from othernetwork cells having high traffic volume). Network traffic (user presences, or average number of active users) classes arethus defined based on the MDT measurements (in the ROPs) in the first zones Z1 andin the union zones Z2^Z3 of the network cells of the cell cluster 130. The ROPsROPA,i (i = 1, 2, …, P) of the first acquisition time interval INTA and the ROPs ROPB,j(j = 1, 2, …, Q) of the second acquisition time interval INTB (and, where defined, alsothe ROPs ROPR,k, k = 1, 2, …, N, of the reference acquisition time interval INTR) areclassified in the proper class, i.e., in the bins BinL or BinH. Then, as described in connection with the first invention embodiment, weightsare calculated, to be assigned to the ROPs ROPB,j (j = 1, 2, …, Q) of the secondacquisition time interval INTB (and, when the reference acquisition time interval INTRis used, also weights to be assigned to the ROPs ROPA,i, i = 1, 2, …, P, of the firstacquisition time interval INTB).The rewarding functions WAand WBfor the two network configurations, before and after the change, are calculated and compared to each other, to assess whether the new network configuration has achieved an improvement in the QoS. Combinations of the two network traffic homogenization methods described in the foregoing are also possible, in which for each class determined in accordance with the first network traffic homogenization method, sub-classes are identified in accordance with the second network traffic homogenization method. ^ ^ ^ While in the invention embodiments described in the foregoing the network performance evaluation metric, or QoS evaluation metric, took into account the user throughput, other indicators can be used for the QoS assessment, in addition to the user throughput. For example, the QoS evaluation metric may also take into account, in addition to the user throughput, parameters providing an indication of the number of handovers, and / or parameters providing an indication of how many base stations ofthe mobile network are “active” or “on”, i.e. ̧not “switched off”: indeed, the higherthe number of “on” base stations the higher the user throughput, but this is achieved at the cost of a greater power consumption of the mobile network; the metric may then be defined in order to apply increasing penalties for increasing numbers of active base stations). In any case, the specific metric defined for the evaluation of networkperformance does not constitute a limitation of the present invention.Naturally, in order to satisfy local and specific requirements, a person skilled in the art may apply to the invention described above many logical and / or physical modifications and alterations. More specifically, although the present invention has been described with a certain degree of particularity with reference to preferred embodiments thereof, it should be understood that various omissions, substitutions and changes in the form and details as well as other embodiments are possible. In particular, different embodiments of the invention may even be practiced without the specific details set forth in the preceding description for providing a more thorough understanding thereof; on the contrary, well-known features may have been omitted or simplified in order not to encumber the description with unnecessary details. Moreover, it is expressly intended that specific elements and / or method steps described in connection with any disclosed embodiment of the invention may be incorporated in any other embodiment. * * * * *

Claims

CLAIMS 1. A computer-implemented method for assessing the change in the Quality of Service, QoS, offered by a mobile communications network, consequent to a change in configuration parameters of one or more network cells (115a) of the mobile communications network (100), the method comprising: -defining a network cell cluster (130) including said one or more network cells(115a) and neighboring network cells (115b) adjacent to said one or more network cells (115a), the network cell cluster being defined in such a way as to correspond toa mobile communications network coverage area that remains essentially constantbefore and after the change in configuration parameters of the one or more networkcells (115a), -acquiring network measurements in respect of an area of the mobilecommunications network corresponding to the network cell cluster (130), wherein saidacquiring network measurements comprises: -acquiring, during a first acquisition time interval (INTA) before saidchange in configuration parameters, first network traffic amount indicators and firstQoS indicators, and -acquiring, during a second acquisition time interval (INTB) after saidchange in configuration parameters, second network traffic amount indicators andsecond QoS indicators, -deriving from the acquired first network traffic amount indicators a firstdistribution of network traffic amount (605A) during the first acquisition time interval (INTA); -deriving from the acquired second network traffic amount indicators a seconddistribution of network traffic amount (605B) during the second acquisition timeinterval (INTB); -obtaining a reference distribution of network traffic amount (605A; 905R) inthe network cell cluster (130) during a reference time interval (INTR; INTA);- calculating a first reward based on said first QoS indicators and calculating asecond reward based on said second QoS indicators; -assessing the change in the QoS based on a comparison between the firstreward and the second reward, wherein said calculating the first reward and the second reward comprises: calculating first weights to be applied to the first QoS indicators and second weights to be applied to the second QoS indicators, said first weights and secondweights being calculated so as to make the first distribution of network traffic amount(605A) and the second distribution of network traffic amount (605B) correspond to the reference distribution of network traffic amount (605A; 905R).

2. The method of claim 1, wherein said first acquisition time interval (INTA)comprises a first succession of first elementary acquisition time intervals (ROPA) and said second acquisition time interval (INTB) comprises a second succession of second elementary acquisition time intervals (ROPB), the first network traffic amount indicators and first QoS indicators being acquired during each first elementaryacquisition time interval (ROPA) of the first succession, and the second network trafficamount indicators and second QoS indicators being acquired during each secondelementary acquisition time interval (ROPB) of the second succession.

3. The method of claim 2, wherein said first distribution of network trafficamount (605A) during the first acquisition time interval (INTA) is a distribution of thefirst elementary acquisition time intervals (ROPA) as a function of the first networktraffic amount indicators, and said second distribution of network traffic amount(605B) during the second acquisition time interval (INTB) is a distribution of the second elementary acquisition time intervals (ROPB) as a function of the first networktraffic amount indicators.

4. The method of any one of the preceding claims, wherein said obtaining areference distribution of network traffic amount (905R) comprises:- defining a reference acquisition time interval (INTR), wherein said referenceacquisition time interval comprises a third succession of third elementary acquisitiontime intervals (ROPR); -acquiring, during each third elementary acquisition time interval (ROPR) ofthe third succession, third network traffic amount indicators and third QoS indicators;- deriving from the third network traffic amount indicators the referencedistribution of network traffic amount (605R), the reference distribution of networktraffic amount being a distribution of the third elementary acquisition time intervals(ROPR) as a function of the third network traffic amount indicators.

5. The method of any of claims 1-3, wherein said reference distribution ofnetwork traffic amount (605A) is the first distribution of network traffic amount (605A) during the first acquisition time interval (INTA), said obtaining a reference distribution of network traffic amount including said deriving from the first network traffic amountindicators the first distribution of network traffic amount (605A) during the firstacquisition time interval (INTA), and wherein said first weights to be applied to the first QoS indicators are unitary weights and said second weights are calculated so asto make the second distribution of network traffic amount (605B) correspond to thefirst distribution of network traffic amount (605A).

6. The method of claim 4 or claim 5, wherein said weights are calculated so as to obtain, from the distribution of the second elementary acquisition time intervals(ROPB) as a function of the first network traffic amount indicators, a re-shapeddistribution having a shape corresponding to one among: the distribution of the thirdelementary acquisition time intervals (ROPR) as a function of the third network trafficamount indicators, the distribution of the first elementary acquisition time intervals(ROPA) as a function of the first network traffic amount indicators.

7. The method of any one of claims 4-6, wherein said calculating weights to be applied to the second network measurements comprises: -defining a plurality of classes of network traffic amount (BinA,1-BinA,5;BinR,1-BinR,5) in respect of the reference network traffic amount distribution, eachclass of network traffic amount of said plurality of classes of network traffic amountcorresponding to a respective reference network traffic amount; -classifying each of the first elementary acquisition time intervals (ROPA) andeach of the second elementary acquisition time intervals (ROPB) in a respective classof said plurality of classes according to the respective first and second network trafficamount indicators;- determining a number of first elementary acquisition time intervals (ROPA)and second elementary acquisition time intervals (ROPB) in each class of said pluralityof classes, and -calculating the weights to be applied to the second network measurementsbased on the determined number of first elementary acquisition time intervals (ROPA)and second elementary acquisition time intervals (ROPB) in each class of said pluralityof classes.

8. The method of any one of the preceding claims, comprising aggregating, over all the network cells of the cell cluster (130), the acquired first network trafficamount indicators and first QoS indicators, and the acquired second network trafficamount indicators and second QoS indicators.

9. The method of any one of the preceding claims, wherein: -said first and second network traffic amount indicators include at least one oftraffic volume measurements, number of active users, presences of users, and -said first and second QoS indicators include indicators of a distribution ofnetwork traffic amounts as a function of a user throughput that the mobilecommunications network is capable of providing to users.

10. The method of any one of the preceding claims, wherein said first andsecond network traffic amount indicators, and said first and second QoS indicators areacquired from the mobile communications network as data aggregated at the level of each network cell of the cell cluster (130).

11. The method of any one of claims 1-9, wherein said first and second networktraffic amount indicators, and said first and second QoS indicators are acquired as geo-referenced data, particularly MDT data, the method further comprising aggregating the acquired geo-referenced data at the level of sub-areas of the coverage area of each network cell of the cell cluster (130).

12. A system for assessing the change in the Quality of Service, QoS, offered by a mobile communications network, consequent to a change in configuration parameters of one or more network cells (115a) of the mobile communications network (100), the system being configured to: -acquire network measurements in respect of an area of the mobilecommunications network corresponding to a network cell cluster (130) including saidone or more network cells (115a) and neighboring network cells (115b) adjacent tosaid one or more network cells (115a), the network cell cluster corresponding to amobile communications network coverage area that remains essentially constant before and after the change in configuration parameters of the one or more network cells (115a), wherein the system is configured to acquire said network measurements by: -acquiring, during a first acquisition time interval (INTA) before saidchange in configuration parameters, first network traffic amount indicators and firstQoS indicators, and -acquiring, during a second acquisition time interval (INTB) after saidchange in configuration parameters, second network traffic amount indicators andsecond QoS indicators; -obtain a reference distribution of network traffic amount (605A; 905R) adaptedto provide a reference distribution of network traffic amount in the network cell cluster (130) during a reference time interval (INTR; INTA); -derive from the first network traffic amount indicators a first distribution ofnetwork traffic amount (605A) during the first acquisition time interval (INTA); -derive from the acquired second network traffic amount indicators a seconddistribution of network traffic amount (605B) during the second acquisition time interval (INTB); -calculate a first reward based on said first QoS indicators and calculating asecond reward based on said second QoS indicators; -assess the change in the QoS based on a comparison between the first rewardand the second reward,wherein the system is configured to calculate the first reward and the second rewardby: calculating first weights to be applied to the first QoS indicators and second weights to be applied to the second QoS indicators, said first weights and second weights being calculated so as to make the first distribution of network traffic amount (605A) and the second distribution of network traffic amount (605B) correspond to the reference distribution of network traffic amount (605A; 905R). * * * * *

Images