A system for generative artificial intelligence-based network configuration simulation and optimization

Abstract

The present invention relates to a system (1) for optimizing network configurations and improving quality of service by using generative adversarial networks (GANs).

Description

[0001] A SYSTEM FOR GENERATIVE ARTIFICIAL INTELLIGENCE-BASED NETWORK CONFIGURATION SIMULATION AND OPTIMIZATION

[0002] Technical Field

[0003] The present invention relates to a system for optimizing network configurations and improving quality of service by using generative adversarial networks (GANs).

[0004] Background of the Invention

[0005] Today, methods and systems for managing the performance of edge network units in edge data networks are available. In particular, a measurement work request generated by an edge consumer and designed to collect performance measurement information associated with at least one edge network unit is enabled to be generated and sent, and this information to be received by an edge producer. This method aims to assess network performance and potentially trigger appropriate actions. Among the edge network units are edge application servers (EAS), edge enabler servers (EES), and edge configuration servers (ECS). This system performs actions in the form of scaling of edge network units according to specific performance criteria by enabling detailed collection and analysis of performance measurement information. However, studies for optimizing network configurations by using Generative Adversarial Networks (GANs), increasing quality of service, indirectly improving network performance through the use of deep learning algorithms, providing network administrators and service providers with the capacity to assess the potential impacts of planned configuration changes in advance, contributing to strategic decision-making process and allowing for longterm network optimization are not common. For this reason, considering the studies and deficiencies included in the current technique, it is understood that there is a need for a system for optimizing network configurations and improving quality of service by using generative adversarial networks (GANs).

[0006] The International patent document no. WO2023211572A1, an application included in the state of the art, discloses a system for the use of artificial intelligence (Al) and machine learning (ML) between the user equipment (UE) and the network in telecommunication systems. The system included in the said invention comprises a plurality of predetermined collaboration levels in order to determine the level of collaboration between the user equipment (UE) and the network and to perform air interface optimization based on this level. The collaboration levels differ according to the type and level of collaboration between the network and the UE. These collaboration levels are realized by the first unit in order to determine the AI / ML collaboration level. Then, based on the determined collaboration level, at least one AI / ML model is enabled to be used for air interface optimization to the UE. This process is carried out by the second unit. The collaboration levels represent the different structures and functions of AI / ML collaboration. At the first level, AI / ML collaboration between the network and the UE is realized based on signaling without model transfer, while at the second level, AI / ML collaboration is realized based on signaling without model transfer. At the third level, AI / ML collaboration based on model transfer and signaling is realized. In addition, AI / ML model training is also performed in the system. This training can take place both in the network and in the UE, and different AI / ML models are used depending on the level of training. Model transfer can be realized by transferring parameters or parts of an existing AI / ML model or by transferring a new AI / ML model. Reports from the UE are used in determining the collaboration level. In addition, the request for information about AI / ML models stored in the UE or to be used is also transmitted by the first unit. The air interface optimization may include functions such as Channel State Information (CSI) feedback enhancement, beam management, and positioning accuracy enhancement. The modules in the system (data collection, model training, model inference and actor) are used as software, hardware or a combination thereof and processor-readable instructions are used to perform the functions of the modules.

[0007] Summary of the Invention

[0008] An object of the present invention is to realize a system developed with the aim of optimizing network configurations and improving quality of service by using generative adversarial networks (GANs).

[0009] Another object of the present invention is to realize a system developed with the aim of enabling network administrators and service providers to assess the potential impacts of planned configuration changes in advance.

[0010] A further object of the present invention is to realize a system developed with the aim of predicting possible quality of service outputs for given configuration changes by using deep learning algorithms obtained by training the GAN model.

[0011] A further object of the present invention is to realize a system developed with the aim of examining and analyzing configuration parameters that directly affect network performance.

[0012] A further object of the present invention is to realize a system developed with the aim of supporting the decision-making process for optimizing network configurations and quality of service; providing a significant advancement in network management and performance improvement; modeling the complexity of network configurations and quality of service through the use of generative adversarial networks and predicting the impacts of future changes, thus, providing network administrators and service providers with a strategic advantage in maximizing the performance of their networks and improving user satisfaction. Detailed Description of the Invention

[0013] “A System for Generative Artificial Intelligence-Based Network Configuration Simulation and Optimization” realized to fulfd the objectives of the present invention is shown in the figures attached, in which:

[0014] Figure 1 is a schematic view of the inventive system.

[0015] Figure 2 is a view of the block diagram of the training of generative artificial intelligence model in the inventive system.

[0016] The components illustrated in the figures are individually numbered, where the numbers refer to the following:

[0017] 1. System

[0018] 2. Data Collection and Processing Server

[0019] 3. Model Development and Training Server

[0020] 4. Simulation and Assessment Server

[0021] 5. Application and Integration Server

[0022] 6. User Interface and Reporting Server

[0023] B. Base Station

[0024] K. User Equipment

[0025] The inventive system (1) developed with the aim of optimizing network configurations and improving quality of service by using generative adversarial networks comprises

[0026] at least one data collection and processing server (2) which is configured to predict the location of the relevant users by analyzing the signal levels received from user equipment (K) based on these data; to perform this process by using spatial analysis and location prediction algorithms; to collect KPI data such as signal strength, data transfer rates and the like received from user equipment (K); to enable these data to be used in the processes of measuring and analyzing quality of service; to convert the collected KPI point data into a grid-based data structure that provides a broader analysis and assessment opportunity; to apply various statistical processes in the form of aggregation, averaging, median calculation with the aim of combining the data during this conversion; to store the obtained grid-based data in a two-dimensional matrix structure, each cell of which expresses quality of service of the relevant region; to collect the location and configuration information in the form of antenna angle, tilt, modulation and frequency band of the base stations (B) serving the region; to use this information in the processes of assessment and improvement of quality of service; to integrate location-based KPI data and configuration information of base stations (B) in a data set; to store KPI data in the two-dimensional grid structure and configuration information in the vector structure, thus, to provide a comprehensive and detailed data set for model development and training processes;

[0027] at least one model development and training server (3) which is configured to perform the training of a machine learning model that learns the relations between network configurations and quality of service and models these relations by using advanced generative artificial intelligence (GAI) techniques such as GAN; to access the data obtained from the data collection and processing server (2) for training the model; to use these data in the training process of the model, thus, to enable the model to learn the relation between quality of service information and the environment that provides this quality of service by processing them; to perform parameter optimization in order to maximize the performance of the model during model training; to perform hyperparameter tuning and fine-tuning processes in order to improve the accuracy and predictive capability of the model in this process; to store the trained models after the training process is completed; to enable the models to be easily accessible and reusable in future analyses or applications by storing them, thus, to allow for continuous improvement and updating of the model, while at the same time providing importance in terms of model management and version check;

[0028] at least one simulation and assessment server (4) which is configured to simulate how the developed model will perform by using real-world data and scenarios; to use these simulations to observe how the model behaves under various conditions and variations; to assess the performance of the model based on simulation results; to analyze the results of the assessment to determine the strengths and areas that need to be improved of the model; to determine the direction of improvements to be made in order to increase the performance of the model through this analysis; to help mitigate the potential risks of the model before application through simulation and assessment processes, thus, to enable network administrators and service providers to make informed decisions before integrating the model into real-world applications; to allow the model to be continuously improved through these processes and to provide more accurate and reliable predictions over time;

[0029] at least one application and integration server (5) which is configured to perform the integration of the developed model with real-time data flows and existing systems; to enable the model to operate continuously with up-to-date data and to perform effectively in dynamic network environments with this integration; to provide the necessary interfaces and application programming interfaces (APIs) in order to enable the model to operate in harmony with existing network management and operational systems, thus, to allow the model to contribute to decision-making and network optimization processes by being directly integrated into network management processes; to perform the updates and maintenance processes required for the developed model to adapt to changing network conditions and needs over time; to continuously monitor the performance of the model, thus, to maintain the accuracy and reliability of the model over time; and to provide interfaces that make it easy for users to interact with the predictions and analysis provided by the model, thus, to enable network administrators and service providers to easily understand the information provided by the model and make informed decisions based on this information; and

[0030] at least one user interface and reporting server (6) which is configured to design and provide user-friendly interfaces that present the predictions and analyses generated by the model; to enable users to easily understand and interpret the data obtained from the model, thus, to enable them to make network management and optimization decisions based on information; to generate detailed reports on network performance, quality of service and other key metrics; to provide in-depth analysis and presentation of data through these reporting tools; to provide users with comprehensive information about network conditions; to provide users with dashboards that allow them to perform interactive analysis on the data presented by the model; to enable users to perform operations in the form of creating their own queries on data sets, exploring different scenarios and assessing data from various perspectives with these interactive tools; to be designed in such a way as to enable network administrators and service providers to make more informed decisions based on the information provided by the model, thus, to increase effectiveness and efficiency in network optimization and management processes; to provide users with mechanisms that enable them to provide feedback on model predictions, analysis results and reports; to enable the model and reporting tools to better serve user needs by improving the user experience through feedback mechanisms; to provide users with reporting and dashboard options that they can customize according to their needs and preferences; to enable users to filter data, choose from different visualization options and highlight the most relevant information according to their specific analysis needs with these customization options; to allow users to use the information obtained from the model more effectively with customizable reporting and indicator options and to facilitate performing more in-depth analysis in network management processes. The data collection and processing server (2) included in the inventive system (1) is configured to collect KPI (Key Performance Indicator) data such as signal strength, data transfer rates and the like received for each location by performing location prediction from the signal levels received by the user equipment (K) of customers; to make a transition to a grid-based data structure that provides a broader perspective than location-based point data and is stored in a two-dimensional matrix structure where quality of service is assessed from the perspective of the user; to apply processes in the form of aggregation, averaging and median in order to combine the data during this conversion; to collect the configuration information in the form of location, antenna angle, tilt, modulation and frequency band of the base stations (B) serving the region; to create an integrated data set with location-based KPI data in two-dimensional grid structure and configuration information in vector structure; and to prepare this integrated data set to be used in subsequent model development and training processes. The data collection and processing server (2) is configured to predict the location of the relevant users by analyzing the signal levels received from user equipment (K) based on these data and to perform this process by using spatial analysis and location prediction algorithms. The data collection and processing server (2) is configured to collect KPI data such as signal strength, data transfer rates and the like received from user equipment (K) and to enable these data to be used in the processes of measuring and analyzing quality of service. The data collection and processing server (2) is configured to convert the collected KPI point data into a grid-based data structure that provides a broader analysis and assessment opportunity and to apply various statistical processes in the form of aggregation, averaging, median calculation with the aim of combining the data during this conversion. The data collection and processing server (2) is configured to store the obtained grid-based data in a two-dimensional matrix structure, each cell of which expresses quality of service of the relevant region. The data collection and processing server (2) is configured to collect the location and configuration information in the form of antenna angle, tilt, modulation and frequency band of the base stations (B) serving the region and to use this information in the processes of assessment and improvement of quality of service. The data collection and processing server (2) is configured to integrate locationbased KPI data and configuration information of base stations (B) in a data set; to store KPI data in the two-dimensional grid structure and configuration information in the vector structure, thus, to provide a comprehensive and detailed data set for model development and training processes.

[0031] The model development and training server (3) included in the inventive system (1) is configured to enable the Generative Adversarial Networks (GAN) model, one of the Generative Artificial Intelligence (GAI) approaches, to be trained; to train the GAN model designed to understand the relation between quality of service and the environment information that provides this quality of service by using the data received from the data collection and processing server (3); to fulfill the responsibilities in the machine learning training stages in the form of parameter optimization and model storage during the training of the model; and to increase the capacity to predict the effects of configuration changes by understanding the relation between quality of service and network configurations in depth. The model development and training server (3) is configured to perform the training of a machine learning model that learns the relations between network configurations and quality of service and models these relations by using advanced generative artificial intelligence (GAI) techniques such as GAN. The model development and training server (3) is configured to access the data obtained from the data collection and processing server (2) for training the model; to use these data in the training process of the model, thus, to enable the model to learn the relation between quality of service information and the environment that provides this quality of service by processing them. The model development and training server (3) is configured to perform parameter optimization in order to maximize the performance of the model during model training and to perform hyperparameter tuning and fine-tuning processes in order to improve the accuracy and predictive capability of the model in this process. The model development and training server (3) is configured to store the trained models after the training process is completed; to enable the models to be easily accessible and reusable in future analyses or applications by storing them, thus, to allow for continuous improvement and updating of the model, while at the same time providing importance in terms of model management and version check.

[0032] The simulation and assessment server (4) included in the inventive system (1) is configured to test and assess the performance of the developed model; to simulate how the model performs under real- world conditions; to enable the model to be assessed for accuracy, precision and other important metrics through these simulations; to perform this assessment process in order to determine the strengths and areas that need improvement of the model; to help the model to be continuously improved and to mitigate risks before application. The simulation and assessment server (4) is configured to simulate how the developed model will perform by using real-world data and scenarios; to use these simulations to observe how the model behaves under various conditions and variations. The simulation and assessment server (4) is configured to assess the performance of the model based on simulation results. The simulation and assessment server (4) is configured to analyze the results of the assessment to determine the strengths and areas that need to be improved of the model and to determine the direction of improvements to be made in order to increase the performance of the model through this analysis. The simulation and assessment server (4) is configured to help mitigate the potential risks of the model before application through simulation and assessment processes, thus, to enable network administrators and service providers to make informed decisions before integrating the model into real- world applications; to allow the model to be continuously improved through these processes and to provide more accurate and reliable predictions over time.

[0033] The application and integration server (5) included in the inventive system (1) is configured to enable the integration of the developed model with end users and other systems; to be responsible for the successful deployment of the model in the application environment; to perform functions in the form of integrating the model into real-world data flows and operational systems, continuously updating and maintaining the model; to be developed to enable network administrators and service providers to continuously use the predictions and analysis provided by the model; and to increase the operational efficiency and usability of the model. The application and integration server (5) is configured to perform the integration of the developed model with real-time data flows and existing systems and to enable the model to operate continuously with up-to-date data and to perform effectively in dynamic network environments with this integration. The application and integration server (5) is configured to provide the necessary interfaces and application programming interfaces (APIs) in order to enable the model to operate in harmony with existing network management and operational systems, thus, to allow the model to contribute to decision-making and network optimization processes by being directly integrated into network management processes. The application and integration server (5) is configured to perform the updates and maintenance processes required for the developed model to adapt to changing network conditions and needs over time; to continuously monitor the performance of the model, thus, to maintain the accuracy and reliability of the model over time. The application and integration server (5) is configured to provide interfaces that make it easy for users to interact with the predictions and analysis provided by the model, thus, to enable network administrators and service providers to easily understand the information provided by the model and make informed decisions based on this information.

[0034] The user interface and reporting server (6) included in the inventive system (1) is configured to be responsible for presenting data obtained from the model to end users in a comprehensible and accessible way; to provide user-friendly interfaces and reporting tools that enable users to interact with model predictions, analyses and quality of service reports; to support informed decision-making by providing network managers and service providers with comprehensive information on network performance and quality of service; to allow users to easily analyze the information provided by the model through interfaces and reporting tools and to optimize network management strategies based on this information. The user interface and reporting server (6) is configured to design and provide user-friendly interfaces that present the predictions and analyses generated by the model; to enable users to easily understand and interpret the data obtained from the model, thus, to enable them to make network management and optimization decisions based on information. The user interface and reporting server (6) is configured to generate detailed reports on network performance, quality of service and other key metrics; to provide in-depth analysis and presentation of data through these reporting tools and to provide users with comprehensive information about network conditions. The user interface and reporting server (6) is configured to provide users with dashboards that allow them to perform interactive analysis on the data presented by the model and to enable users to perform operations in the form of creating their own queries on data sets, exploring different scenarios and assessing data from various perspectives with these interactive tools. The user interface and reporting server (6) is configured to be designed in such a way as to enable network administrators and service providers to make more informed decisions based on the information provided by the model, thus, to increase effectiveness and efficiency in network optimization and management processes. The user interface and reporting server (6) is configured to provide users with mechanisms that enable them to provide feedback on model predictions, analysis results and reports; to enable the model and reporting tools to better serve user needs by improving the user experience through feedback mechanisms. The user interface and reporting server (6) is configured to provide users with reporting and dashboard options that they can customize according to their needs and preferences; to enable users to filter data, choose from different visualization options and highlight the most relevant information according to their specific analysis needs with these customization options; to allow users to use the information obtained from the model more effectively with customizable reporting and indicator options and to facilitate performing more in-depth analysis in network management processes.

[0035] Industrial Application of the Invention In the inventive system (1), network configuration and quality of service data are collected, processed and prepared for training the GAN model. Models that predict the impact of configuration changes on quality of service are developed and trained with the use of these data. While the performance of the model is assessed on real-world data, the model is made accessible to end-users and integrated into units. Model predictions and analysis results are made accessible to users, thus, network administrators and service providers are enabled to make informed decisions on maximizing network performance and increasing user satisfaction.

[0036] Within these basic concepts; it is possible to develop various embodiments of the inventive “A System (1) for Generative Artificial Intelligence-Based Network Configuration Simulation and Optimization”; the invention cannot be limited to examples disclosed herein and it is essentially according to claims.

Claims

CLAIMS1. A system (1) developed with the aim of optimizing network configurations and improving quality of service by using generative adversarial networks; comprisingat least one data collection and processing server (2) which is configured to predict the location of the relevant users by analyzing the signal levels received from user equipment (K) based on these data; to perform this process by using spatial analysis and location prediction algorithms; to collect KPI data such as signal strength, data transfer rates and the like received from user equipment (K); to enable these data to be used in the processes of measuring and analyzing quality of service; to convert the collected KPI point data into a grid-based data structure that provides a broader analysis and assessment opportunity; to apply various statistical processes in the form of aggregation, averaging, median calculation with the aim of combining the data during this conversion; to store the obtained grid-based data in a two-dimensional matrix structure, each cell of which expresses quality of service of the relevant region; to collect the location and configuration information in the form of antenna angle, tilt, modulation and frequency band of the base stations (B) serving the region; to use this information in the processes of assessment and improvement of quality of service; to integrate location-based KPI data and configuration information of base stations (B) in a data set; to store KPI data in the two-dimensional grid structure and configuration information in the vector structure, thus, to provide a comprehensive and detailed data set for model development and training processes; and characterized byat least one model development and training server (3) which is configured to perform the training of a machine learning model that learns the relations between network configurations and quality of service and models these relations by using advanced generative artificial intelligence (GAI) techniques such as GAN; to access the data obtained from the datacollection and processing server (2) for training the model; to use these data in the training process of the model, thus, to enable the model to learn the relation between quality of service information and the environment that provides this quality of service by processing them; to perform parameter optimization in order to maximize the performance of the model during model training; to perform hyperparameter tuning and fine-tuning processes in order to improve the accuracy and predictive capability of the model in this process; to store the trained models after the training process is completed; to enable the models to be easily accessible and reusable in future analyses or applications by storing them, thus, to allow for continuous improvement and updating of the model, while at the same time providing importance in terms of model management and version check;at least one simulation and assessment server (4) which is configured to simulate how the developed model will perform by using real-world data and scenarios; to use these simulations to observe how the model behaves under various conditions and variations; to assess the performance of the model based on simulation results; to analyze the results of the assessment to determine the strengths and areas that need to be improved of the model; to determine the direction of improvements to be made in order to increase the performance of the model through this analysis; to help mitigate the potential risks of the model before application through simulation and assessment processes, thus, to enable network administrators and service providers to make informed decisions before integrating the model into real-world applications; to allow the model to be continuously improved through these processes and to provide more accurate and reliable predictions over time;at least one application and integration server (5) which is configured to perform the integration of the developed model with real-time data flows and existing systems; to enable the model to operate continuously with up-to-date data and to perform effectively in dynamic network environments with this integration; to provide the necessary interfaces and applicationprogramming interfaces (APIs) in order to enable the model to operate in harmony with existing network management and operational systems, thus, to allow the model to contribute to decision-making and network optimization processes by being directly integrated into network management processes; to perform the updates and maintenance processes required for the developed model to adapt to changing network conditions and needs over time; to continuously monitor the performance of the model, thus, to maintain the accuracy and reliability of the model over time; and to provide interfaces that make it easy for users to interact with the predictions and analysis provided by the model, thus, to enable network administrators and service providers to easily understand the information provided by the model and make informed decisions based on this information; andat least one user interface and reporting server (6) which is configured to design and provide user-friendly interfaces that present the predictions and analyses generated by the model; to enable users to easily understand and interpret the data obtained from the model, thus, to enable them to make network management and optimization decisions based on information; to generate detailed reports on network performance, quality of service and other key metrics; to provide in-depth analysis and presentation of data through these reporting tools; to provide users with comprehensive information about network conditions; to provide users with dashboards that allow them to perform interactive analysis on the data presented by the model; to enable users to perform operations in the form of creating their own queries on data sets, exploring different scenarios and assessing data from various perspectives with these interactive tools; to be designed in such a way as to enable network administrators and service providers to make more informed decisions based on the information provided by the model, thus, to increase effectiveness and efficiency in network optimization and management processes; to provide users with mechanisms that enable them to provide feedback on model predictions, analysis results and reports; to enable the model and reporting tools to betterserve user needs by improving the user experience through feedback mechanisms; to provide users with reporting and dashboard options that they can customize according to their needs and preferences; to enable users to filter data, choose from different visualization options and highlight the most relevant information according to their specific analysis needs with these customization options; to allow users to use the information obtained from the model more effectively with customizable reporting and indicator options and to facilitate performing more in-depth analysis in network management processes.

2. A system (1) according to Claim 1; characterized by the data collection and processing server (2) which is configured to collect KPI (Key Performance Indicator) data such as signal strength, data transfer rates and the like received for each location by performing location prediction from the signal levels received by the user equipment (K) of customers; to make a transition to a grid-based data structure that provides a broader perspective than location-based point data and is stored in a two-dimensional matrix structure where quality of service is assessed from the perspective of the user; to apply processes in the form of aggregation, averaging and median in order to combine the data during this conversion; to collect the configuration information in the form of location, antenna angle, tilt, modulation and frequency band of the base stations (B) serving the region; to create an integrated data set with location-based KPI data in two-dimensional grid structure and configuration information in vector structure; and to prepare this integrated data set to be used in subsequent model development and training processes.

3. A system (1) according to Claim 1 or 2; characterized by the data collection and processing server (2) which is configured to predict the location of the relevant users by analyzing the signal levels received from user equipment (K) based on these data and to perform this process by using spatial analysis and location prediction algorithms.

4. A system (1) according to Claim 3; characterized by the data collection and processing server (2) which is configured to collect KPI data such as signal strength, data transfer rates and the like received from user equipment (K) and to enable these data to be used in the processes of measuring and analyzing quality of service.

5. A system (1) according to any one of the preceding claims; characterized by the data collection and processing server (2) which is configured to convert the collected KPI point data into a grid-based data structure that provides a broader analysis and assessment opportunity and to apply various statistical processes in the form of aggregation, averaging, median calculation with the aim of combining the data during this conversion.

6. A system (1) according to any one of the preceding claims; characterized by the data collection and processing server (2) which is configured to store the obtained grid-based data in a two-dimensional matrix structure, each cell of which expresses quality of service of the relevant region.

7. A system (1) according to any one of the preceding claims; characterized by the data collection and processing server (2) which is configured to collect the location and configuration information in the form of antenna angle, tilt, modulation and frequency band of the base stations (B) serving the region and to use this information in the processes of assessment and improvement of quality of service.

8. A system (1) according to any one of the preceding claims; characterized by the data collection and processing server (2) which is configured to integrate location-based KPI data and configuration information of base stations (B) in a data set; to store KPI data in the two-dimensional grid structure and configuration information in the vector structure, thus, to provide a comprehensive and detailed data set for model development and training processes.

9. A system (1) according to any one of the preceding claims; characterized by the model development and training server (3) which is configured to enable the Generative Adversarial Networks model, one of the Generative Artificial Intelligence approaches, to be trained; to train the GAN model designed to understand the relation between quality of service and the environment information that provides this quality of service by using the data received from the data collection and processing server (3); to fulfill the responsibilities in the machine learning training stages in the form of parameter optimization and model storage during the training of the model; and to increase the capacity to predict the effects of configuration changes by understanding the relation between quality of service and network configurations in depth.

10. A system (1) according to any one of the preceding claims; characterized by the model development and training server (3) which is configured to perform the training of a machine learning model that learns the relations between network configurations and quality of service and models these relations by using advanced generative artificial intelligence (GAI) techniques such as GAN.

11. A system (1) according to any one of the preceding claims; characterized by the model development and training server (3) which is configured to access the data obtained from the data collection and processing server (2) for training the model; to use these data in the training process of the model, thus, to enable the model to learn the relation between quality of service information and the environment that provides this quality of service by processing them.

12. A system (1) according to any one of the preceding claims; characterized by the model development and training server (3) which is configured to perform parameter optimization in order to maximize the performance of the model during model training and to perform hyperparameter tuning and fine-tuning processes in order to improve the accuracy and predictive capability of the model in this process.

13. A system (1) according to any one of the preceding claims; characterized by the model development and training server (3) which is configured to store the trained models after the training process is completed; to enable the models to be easily accessible and reusable in future analyses or applications by storing them, thus, to allow for continuous improvement and updating of the model, while at the same time providing importance in terms of model management and version check.

14. A system (1) according to any one of the preceding claims; characterized by the simulation and assessment server (4) which is configured to test and assess the performance of the developed model; to simulate how the model performs under real-world conditions; to enable the model to be assessed for accuracy, precision and other important metrics through these simulations; to perform this assessment process in order to determine the strengths and areas that need improvement of the model; to help the model to be continuously improved and to mitigate risks before application.

15. A system (1) according to any one of the preceding claims; characterized by the simulation and assessment server (4) which is configured to simulate how the developed model will perform by using real-world data and scenarios; to use these simulations to observe how the model behaves under various conditions and variations.

16. A system (1) according to any one of the preceding claims; characterized by the simulation and assessment server (4) which is configured to assess the performance of the model based on simulation results.

17. A system (1) according to any one of the preceding claims; characterized by the simulation and assessment server (4) which is configured to analyze the results of the assessment to determine the strengths and areas that need to beimproved of the model and to determine the direction of improvements to be made in order to increase the performance of the model through this analysis.

18. A system (1) according to any one of the preceding claims; characterized by the simulation and assessment server (4) which is configured to help mitigate the potential risks of the model before application through simulation and assessment processes, thus, to enable network administrators and service providers to make informed decisions before integrating the model into real-world applications; to allow the model to be continuously improved through these processes and to provide more accurate and reliable predictions over time.

19. A system (1) according to any one of the preceding claims; characterized by the application and integration server (5) which is configured to enable the integration of the developed model with end users and other systems; to be responsible for the successful deployment of the model in the application environment; to perform functions in the form of integrating the model into real-world data flows and operational systems, continuously updating and maintaining the model; to be developed to enable network administrators and service providers to continuously use the predictions and analysis provided by the model; and to increase the operational efficiency and usability of the model.

20. A system (1) according to any one of the preceding claims; characterized by the application and integration server (5) which is configured to perform the integration of the developed model with real-time data flows and existing systems and to enable the model to operate continuously with up-to-date data and to perform effectively in dynamic network environments with this integration.

21. A system (1) according to any one of the preceding claims; characterized by the application and integration server (5) which is configured to provide the necessary interfaces and application programming interfaces (APIs) in order to enable the model to operate in harmony with existing network management andoperational systems, thus, to allow the model to contribute to decision-making and network optimization processes by being directly integrated into network management processes.

22. A system (1) according to any one of the preceding claims; characterized by the application and integration server (5) which is configured to perform the updates and maintenance processes required for the developed model to adapt to changing network conditions and needs over time; to continuously monitor the performance of the model, thus, to maintain the accuracy and reliability of the model over time.

23. A system (1) according to any one of the preceding claims; characterized by the application and integration server (5) which is configured to provide interfaces that make it easy for users to interact with the predictions and analysis provided by the model, thus, to enable network administrators and service providers to easily understand the information provided by the model and make informed decisions based on this information.

24. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured to be responsible for presenting data obtained from the model to end users in a comprehensible and accessible way; to provide user-friendly interfaces and reporting tools that enable users to interact with model predictions, analyses and quality of service reports; to support informed decision-making by providing network managers and service providers with comprehensive information on network performance and quality of service; to allow users to easily analyze the information provided by the model through interfaces and reporting tools and to optimize network management strategies based on this information.

25. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured is configured todesign and provide user-friendly interfaces that present the predictions and analyses generated by the model; to enable users to easily understand and interpret the data obtained from the model, thus, to enable them to make network management and optimization decisions based on information.

26. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured is configured to generate detailed reports on network performance, quality of service and other key metrics; to provide in-depth analysis and presentation of data through these reporting tools and to provide users with comprehensive information about network conditions.

27. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured is configured to provide users with dashboards that allow them to perform interactive analysis on the data presented by the model and to enable users to perform operations in the form of creating their own queries on data sets, exploring different scenarios and assessing data from various perspectives with these interactive tools.

28. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured is configured to be designed in such a way as to enable network administrators and service providers to make more informed decisions based on the information provided by the model, thus, to increase effectiveness and efficiency in network optimization and management processes.

29. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured is configured to provide users with mechanisms that enable them to provide feedback on model predictions, analysis results and reports; to enable the model and reporting tools tobetter serve user needs by improving the user experience through feedback mechanisms.

30. A system (1) according to any one of the preceding claims; characterized by the user interface and reporting server (6) which is configured is configured to provide users with reporting and dashboard options that they can customize according to their needs and preferences; to enable users to filter data, choose from different visualization options and highlight the most relevant information according to their specific analysis needs with these customization options; to allow users to use the information obtained from the model more effectively with customizable reporting and indicator options and to facilitate performing more in-depth analysis in network management processes.

Images