Experience-Oriented Adaptive Resource Management System in Mobile Networks

Abstract

The invention is an integrated system in the mobile communications sector that manages and autonomously optimizes limited network resources (radio resource blocks, power, spectrum) using machine learning methods based on both technical network metrics and subscriber perception-based experience data. The system combines individual subscriber-based technical data (throughput, drop rate, latency) and historical survey scores to generate a dynamic user experience score (QoE) for each user and adaptively allocates network resources based on this score. Thus, the system proactively improves the overall customer experience and minimizes customer churn by prioritizing subscribers at risk of dissatisfaction or with low fault tolerance.

Description

1 TARIFF Experience-Oriented Adaptive Resource Management System in Mobile Networks Technical Area 5 The invention has implications for the telecommunications and mobile communications sectors, particularly for mobile operators' networks. resources (bandwidth, priority, frequency spectrum, etc.) based on subscriber-based experience scores Optimizing in real-time according to Quality of Education (QoE); artificial intelligence using survey data and technical network metrics. Advanced analytics 10 offers a personalized resource management infrastructure by blending it with algorithms. and is used in the field of decision support systems. State of the Art Mobile communication service providers have limited radio 15 broadcasts on thousands of base stations. Efficient use of resources is of critical importance. Existing resource management systems Radio Resource Management (RRM) typically uses similar static rules for all subscribers, or It assigns resources based solely on the instantaneous signal level. However, this approach... the subscriber's current perception of their experience (for example, a user who has experienced repeated problems has a low rating) It ignores the tolerance threshold and past satisfaction data. This situation is particularly problematic in the 20 Quality of service (QoS) technically appears good for "risky" or "vulnerable" subscriber groups. However, this leads to low customer satisfaction (QoE). Radio Resource Management in Mobile Communications Networks And user experience improvement processes are traditionally carried out using these two methods: 25 Static Network Parameters: General parameters defined for all subscribers or subscriber groups. Priority rules (QoS Class Identifiers - QCI), Reactive Customer Complaint Management: Regional processes implemented after a subscriber experiences a problem and files a complaint. Technical reviews and improvements. 30 However, these methods are insufficient for 4G / 5G network infrastructures where user expectations are increasing. The current systems remain; the technical KPIs (throughput, latency, etc.) in a cell are normal. It can report that it is within the limits, but the current technical level is a certain It cannot measure whether the subscriber is satisfied with their "perceived experience" (QoE). 35 2 These shortcomings can be summarized as follows: Ignoring User Perception: Current applications only allocate resources when considering user perception. Focusing on the current signal quality (RSRP / RSRQ); the connection experience the subscriber has had in the past. taking into account the problems, the dissatisfaction expressed in surveys, and the "low error tolerance" It does not include. 5 "One-Size-Fits-All" Approach: Network resources are tailored to the subscriber's actual needs at that moment. or are distributed according to standard rules regardless of past experience points. This situation, experiencing a lack of resources at a critical moment (for example, during an important video call) This leads to the loss (churn) of a sensitive subscriber. 10 Reactive Intervention Structure: Improvement processes only occur when a subscriber becomes unhappy and leaves the system. It starts when the customer reaches that point or opens a complaint record. Technical data with the customer. A proactive resource shift because a "live" link has not been established between satisfaction and the lack of a connection. It cannot be done. 15 Analytical Data Disconnect: Operators have a gap between their vast "subscriber survey data" and "real-time" data. Network traffic data is stored on different platforms. These two datasets can be combined using machine learning. There is no integration that can blend resources and make instant resource decisions. Resource Efficiency Loss: Network resources are only allocated to subscribers with high signal strength. When focused on, it shows a customer using a critical service with a moderate signal level at that moment, and satisfaction. Subscribers with low thresholds are being disadvantaged, which harms overall customer loyalty. It provides. In conclusion, current practices mean that technical network capacity depends on the subscriber's individual experience. There is no system that instantly and adaptively optimizes based on the score. This invention, the word By addressing the gap in the subject matter, it combines technical metrics with "subscriber psychology and perception," and is a learning tool. and a unique infrastructure that allocates resources in a way that achieves the highest "satisfaction yield". It offers. 30 Purpose of the Invention The invention is created by drawing inspiration from existing situations and aims to solve the aforementioned drawbacks. It aims to... 35 3 The main purpose of the invention is to utilize the limited radio resources (spectrum, (power, coding, etc.) dynamically designed to provide the highest quality of customer satisfaction (QoE). managing, synchronizing subscriber perception with technical performance, and preventing operators from losing subscribers. Our goal is to offer a new generation of adaptive management systems that proactively prevent churn (problems / shuttering). Another aim of the invention is to be based solely on signal quality, unlike existing techniques. Instead of making static resource assignments, a "personalized" system is created based on the subscriber's real-time experience score. a system that "prioritizes" and can learn the subscriber's tolerance thresholds over time to present. Another aim of the invention is to analyze technical KPIs from network layers and subscribers' data. By combining satisfaction data obtained from surveys for the first time on the same analytical level, technically... Identifying micro-deviations that appear "normal" but create "unhappiness" on a subscriber basis. The goal is to generate a live "Experience Score" for each subscriber. Another purpose of the invention is to promote reinforcement learning and deep learning. By using algorithms, it can determine which subscriber needs additional resources (Resource Block) at which time. it heard, and guessed based on similar user behavior the system had encountered before. By doing so, resources are not only given to those "in need," but also to those "who will leave the system if their needs are not met." The aim is to ensure that they are directed towards "the one with the highest risk". 20 Another aim of the invention is to target the "minority" who do not complain despite experiencing technical problems on the network. The group is automatically identified based on low experience scores, and these subscribers are redirected to "high" experience scores. By marking them as "priority," they will receive a more stable and higher-level response in their next connection requests. It is about proactively increasing commitment by allocating quality resources. 25 Another purpose of the invention is to reduce resources using a linear logic when network congestion increases. Instead, by protecting subscribers whose experience score has fallen below the critical threshold, it offers limited coverage. capacity with a "smart scale" to maximize overall customer satisfaction. to distribute. 30 Another purpose of the invention is to reduce the reliance of planning engineers on manual parameter changes. a new campaign or a new subscriber profile (e.g., increased video content consumption) When input into the system, the algorithm learns changing experience expectations on its own, using the resource. The aim is to provide a structure that updates the management logic. 35 4 Another purpose of the invention is to help marketing teams identify which "Experience Risk Maps" are suitable for their needs. by enabling users to see which subscriber groups in the region carry "network-related risks," By ensuring that technical investments are directly focused on "customer retention," a return on investment is achieved. The goal is to accelerate the return on investment (ROI). Another aim of the invention is to leverage the "Network Slicing" capabilities offered by 5G for subscribers. The goal is to lay the groundwork for a "personalized slicing" infrastructure by combining it with experience-based data. Another aim of the invention is to treat network resources not just as an engineering asset, but as a... positioning resource management as a "customer satisfaction tool" and transforming it from a reactive process to a 10 The goal is to offer a unique system architecture that transforms the process into a proactive and subscriber-centric one. Figures that will help understand the invention. Figure 1 shows the general architecture of the system that is the subject of the invention. Explanation of Part References 1. Survey and Subscriber Experience Database 2. Network Technical KPI Pool 3. Quality of Experience (QoE) Calculation Module 4.ML-Based Resource Optimization Engine 20 5. Adaptive Radio Source Assignor 6. Experience Monitoring and Feedback Unit Detailed Description of the Invention This detailed description of the preferred configurations of the invention provides a better understanding of the subject matter. The invention, aimed at facilitating understanding and without imposing any limiting effects, concerns mobile communication. In the sector, the limited network resources (radio source blocks, power, spectrum) are utilized both technically. Machine learning based on both network metrics and subscriber perception-based experience data. It is an integrated system that manages and autonomously optimizes through various methods. The system consists of 30 individual components. Combines subscriber-based technical data (throughput, drop rate, latency) and past survey scores. By doing so, it generates a dynamic experience score (QoE) for each user and allocates network resources. It distributes the results adaptively based on this score. Thus, the system minimizes the risk of dissatisfaction. or by proactively prioritizing subscribers with low fault tolerance, thereby improving the overall customer experience. This improves the process and minimizes customer churn. 35 Invention; Survey and Subscriber Experience Database (1), Network Technical KPI Pool (2), Experience The Quality of Efficiency (QoE) Calculation Module (3), ML-based Resource Optimization Engine (4), Adaptive From the components of Radio Source Assigner (5) and Experience Monitoring and Feedback Unit (6) It consists of. The system primarily uses the subscriber's past through the Survey and Subscriber Experience Database (1). satisfaction trends and the subscriber's current status via the Network Technical KPI Pool (2) It collects connection quality data in real-time. These heterogeneous datasets form the Experience Score. (QoE) is processed through the Calculation Module (3) to determine the subscriber's current technical needs and psychological needs. A mathematical experience score is derived by combining expectations. 10 ML Based Resource Optimization Engine (4), these prepared scores and the instantaneous network occupancy It analyzes the ratios. Various algorithms (e.g., Reinforcement Learning) By using (Learning), it can be determined to which subscriber which proportion of limited network resources should be allocated. decides. After the diagnosis and decision phase, the Adaptive Radio Source Assignor (5) physical 15 by dynamically reprogramming resource blocks on the layer, It temporarily assigns high priority to subscribers with low scores. The results of this intervention are monitored through the Experience Monitoring and Feedback Unit (6). The amount of improvement in the subscriber's experience is fed back into the system by Optimization Engine 20. The decision-making mechanism of (4) is constantly updated. Thanks to this cycle, the system adapts to different subscribers. By learning over time how their profiles respond to different technical improvements, they become more accurate sources. They start making decisions. Thus, the invention utilizes network resources not through a static engineering approach, but through a subscriber experience 25 by managing it like a focused, living organism; enabling operators to achieve maximum with limited capacity. This enables it to ensure customer satisfaction and loyalty.

Claims

6 REQUESTS 1. Limited radio resources (spectrum, power, encoding, etc.) in mobile communication networks. dynamically managing subscribers to ensure the highest quality of customer satisfaction (QoE). Synchronizing perception with technical performance and proactively enabling operators to prevent subscriber churn. It is a new generation adaptive management system that prevents such problems, and its feature is; • Surveys that record subscriber satisfaction scores, complaint history, and perceptual data. Subscriber Experience Database (1), • Network Technical KPIs, which collect cell-based traffic, drop rate, and connection quality metrics. Pool (2), 10 • Generating an instant satisfaction score for each subscriber by blending perceptual and technical data. Experience Score (QoE) Calculation Module (3), • Reinforcement Learning determines which subscriber should be given priority for resource (RB) assignment. ML Based Resource Optimization Engine (4), • 15 that dynamically performs bandwidth, frequency, and power assignments at the physical layer Adaptive Radio Source Assigner (5), • measures the impact of resource allocation on the subscriber's experience score and trains the system. Experience Monitoring and Feedback Unit (6) It includes.

2. The system is compliant with Request 1 and its feature is; via Survey and Subscriber Experience Database (1). This involves the subscriber incorporating their past satisfaction trends and tolerance thresholds into the system.

3. The system compliant with Request 1 is characterized by; simultaneously Network Technical KPI Pool (2) This involves collecting the subscriber's current connection quality and signal level metrics. 25 4. It is a system that complies with Request 1 and its feature is; Experience Score (QoE) Calculation Module (3) By correlating survey scores with technical metrics, we can determine the subscriber's current "hunger for experience". and it scores the risk level.

5. It is a system that complies with Request 1 and its feature is; ML Based Resource Optimization Engine (4) By analyzing the scores calculated through this method and network density, the most efficient use of resources is determined. It is about creating a (happiness-focused) distribution plan. 7 6. It is a system that complies with Request 1 and its feature is determined via the Adaptive Radio Source Assignor (5). By communicating the plan to physical network elements, temporary high priority is given to low-scoring subscribers. This is a (prioritization) assignment.

7. The system complies with Request 1 and its feature is; Experience Monitoring and Feedback Unit (6) via 5 Measuring post-assignment improvement and adding the achieved success to the Optimization Engine (4) It is transmitted as training data.

Images