# US 2025/0168669 A1 · System and Method for Intelligently Analyzing and Applying Wireless Communication Network Knowledge Graph

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0168669 A1
- Document ID: US20250168669A1
- Status: Published application
- Kind: A1
- Application number: 18/839,424
- Filed: 2022-11-01
- Published: 2025-05-22
- Earliest claimed: 2022-02-17
- Assignees: Purple Mountain Laboratories
- Applicants: Purple Mountain Laboratories
- Inventors: Shiwen HE; Yongming HUANG; Peng REN; Hang ZHAN; Liangpeng WANG; Zhenyu AN; Yunshan YI; Xiaohu YOU
- Web page: https://patentgrep.com/patent/US20250168669A1

## CPC

- H04W 24/08: Testing, {supervising or monitoring} using real traffic
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 3/00: Computing arrangements based on biological models
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- H04W 24/02: Arrangements for optimising operational condition
- H04W 24/04: Arrangements for maintaining operational condition
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Provided are a system and method for intelligently analyzing and applying a wireless communication network knowledge graph. The system includes at least a knowledge graph unit, an intelligent traceability unit, and a tuning strategy unit. The knowledge graph unit is configured to construct the wireless communication network knowledge graph, and achieve deep analysis and cyclic tuning of the graph in combination with wireless communication network data; the intelligent traceability unit is configured to detect a network anomaly and trace a cause and a source; and the tuning strategy unit is configured to generate a plurality of tuning strategies, determine an optimal tuning strategy, and perform tuning on a wireless communication network. According to the system provided in the present disclosure, intelligent closed-loop feedback is formed, and cyclic tuning is performed on a wireless communication network, thereby effectively improving the performance of the wireless communication network and QoE of users.

## Claims

### Claim at position 1

1-15. (canceled)

### Claim 16

A method for intelligently analyzing and applying a wireless communication network knowledge graph, comprising:
constructing the wireless communication network knowledge graph based on endogenous factors of a wireless communication network and a graph construction method, and performing deep analysis and graph cyclic tuning in combination with classified wireless communication network data;
determining a diagnostic positioning result for a wireless communication network anomaly based on the tuned wireless communication network knowledge graph, an anomaly detection algorithm and an intelligent reasoning algorithm; and
determining a plurality of tuning strategies based on a strategy generation algorithm and the diagnostic positioning result, and determining, in combination with an intelligent decision algorithm, one optimal tuning strategy from the plurality of tuning strategies by using a best result when considering an execution efficiency and an execution effect of a tuning strategy as a target, so as to perform tuning on the wireless communication network.

### Claim 17 (depends on 16)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 16, wherein constructing the wireless communication network knowledge graph based on the endogenous factors of the wireless communication network and the graph construction method, and performing deep analysis and graph cyclic tuning in combination with the classified wireless communication network data comprises:
determining new relationships among various entities in the wireless communication network knowledge graph, analyzing association degrees of the relationships among the various entities, and updating an inter-entity relationship library based on a graph model analysis algorithm and a data model analysis algorithm; and/or determining one or more of:
new entities, new entity types, and new entity attributes in the wireless communication network knowledge graph, and updating an entity library; and
updating the wireless communication network knowledge graph based on the updated entity library and/or the updated inter-entity relationship library, wherein
the entity library and the inter-entity relationship library are determined based on the endogenous factors of the wireless communication network; and the endogenous factors of the wireless communication network comprises indicators and data fields specified by wireless communication network protocols.

### Claim 18 (depends on 16)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 16, wherein determining the diagnostic positioning result for the wireless communication network anomaly based on the tuned wireless communication network knowledge graph, the anomaly detection algorithm and the intelligent reasoning algorithm comprises:
performing anomaly detection on the wireless communication network based on the tuned wireless communication network knowledge graph and the anomaly detection algorithm, so as to determine a fault type; and
determining, based on the fault type and the intelligent reasoning algorithm, one or more entities to which the fault type belongs in the wireless communication network knowledge graph.

### Claim 19 (depends on 16)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 16, wherein determining the plurality of tuning strategies based on the strategy generation algorithm and the diagnostic positioning result, and determining, in combination with the intelligent decision algorithm, the one optimal tuning strategy from the plurality of tuning strategies by using the best result when considering the execution efficiency and the execution effect of the tuning strategy as the target, so as to perform tuning on the wireless communication network comprises:
determining the plurality of tuning strategies of the wireless communication network based on one or more entities to which a fault type belongs in the wireless communication network knowledge graph, and the strategy generation algorithm; and
based on the intelligent decision algorithm, determining the one optimal tuning strategy by using the best result when considering the execution efficiency and the execution effect of the tuning strategy as the target, so as to perform tuning on the wireless communication network.

### Claim 20 (depends on 16)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 16, the method further comprising:
determining, based on the tuned wireless communication network in combination with a tuning strategy and intelligent algorithm evaluation model, improvement degrees on a network-level performance and a user-level performance of the wireless communication network by the one optimal tuning strategy, and a comprehensive capability evaluation result of each intelligent algorithm used.

### Claim 21 (depends on 20)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 20, wherein determining, based on the tuned wireless communication network in combination with the tuning strategy and intelligent algorithm evaluation model, the improvement degrees on the network-level performance and the user-level performance of the wireless communication network by the tuning strategies, and the comprehensive capability evaluation result of the each intelligent algorithm used comprises:
acquiring data of the tuned wireless communication network;
determining the improvement degrees on a performance of the wireless communication network by the one optimal tuning strategy in two dimensions of a network-level performance evaluation and a user-level performance evaluation; and
based on the improvement degrees and the comprehensive capability evaluation result of each intelligent algorithm, determining an intelligent algorithm that needs to be adjusted, wherein the intelligent algorithm that needs to be adjusted comprises one or more of: a deep analysis algorithm, the anomaly detection algorithm, the intelligent reasoning algorithm, the strategy generation algorithm, and the intelligent decision algorithm.

### Claim 22 (depends on 16)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 16, wherein before constructing the wireless communication network knowledge graph based on the endogenous factors of the wireless communication network and the graph construction method, and performing deep analysis and graph cyclic tuning in combination with the classified wireless communication network data, the method further comprises:
acquiring raw data of the wireless communication network, performing pre-processing on the raw data, and then performing classification storage on the preprocessed data according to different data types, wherein the classification storage of different data types is implemented based on a distributed system architecture.

### Claim 23 (depends on 22)

The intelligent analysis and application method for a wireless communication network knowledge graph as claimed in claim 22, wherein a data type of the different data types comprises one or more of: terminal-side wireless air interface data, base station-side wireless air interface data, core network data, and network management data.

### Claim 24 (depends on 17)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 17, wherein entity types of entities in the entity library comprise one or more of: a network-level performance evaluation indicator, a user-level performance evaluation indicator, a general non-tunable data parameter, and a tunable data parameter; entity relationships in the inter-entity relationship library comprise one or more of: a causal relationship, an implicit relationship, and an explicit relationship.

### Claim 25 (depends on 17)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 17, wherein determining the new relationships among various entities in the wireless communication network knowledge graph, analyzing the association degrees of the relationships among the various entities, and updating the inter-entity relationship library based on the graph model analysis algorithm and the data model analysis algorithm, and/or determining one or more of: new entities, new entity types, and new entity attributes in the wireless communication network knowledge graph, and updating the entity library comprises:
based on the graph model algorithm, determining quantitative measurements of the association degrees of the relationships among the various entities in the wireless communication network knowledge graph, and updating the inter-entity relationship library; and
based on the acquired classified wireless communication network data, determining one or more of: the new entities, the new entity types, the new entity attributes, and the new relationships among the various entities in the wireless communication network knowledge graph, and updating the entity library and/or the inter-entity relationship library.

### Claim 26 (depends on 17)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 17, wherein updating the wireless communication network knowledge graph based on the updated entity library and/or the updated inter-entity relationship library comprises:
determining parameters to be tuned based on wireless communication network protocol specifications and theoretical analysis algorithms, and the updated entity library and/or inter-entity relationship library, wherein the parameters to be tuned in the wireless communication network knowledge graph comprise one or more of: parameters of entities, parameters of entity attributes, and parameters of inter-entity relationships; and
updating the wireless communication network knowledge graph based on the parameters to be tuned.

### Claim 27 (depends on 18)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 18, wherein the fault type comprises a fault corresponding to different entities in the entity library and a fault corresponding to different inter-entity relationships in the inter-entity relationship library.

### Claim 28 (depends on 18)

The intelligent analysis and application method for the wireless communication network knowledge graph as claimed in claim 18, wherein determining, based on the fault type and the intelligent reasoning algorithm, the one or more entities to which the fault type belongs in the wireless communication network knowledge graph comprises:
determining one or more entities to which the fault type belongs, wherein the entity is a specific element in the entity library; and
determining a position, a tunability, and an influence factor of the entity based on an attribute of the entity, wherein the position comprises a specific communication layer in which the entity is located or a position at which the entity is located in a graph structure, and the influence factor is an influence level of other entities on the entity, wherein the other entities are entities that have inter-entity relationships with the entity.

### Claim 29 (depends on 16)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 16, wherein the intelligent decision algorithm comprises one or more of: a Deep Reinforcement Learning (DRL) algorithm, a Markov Decision Process (MDP) algorithm, a biology imitation algorithm, and a statistical learning algorithm.

### Claim 30 (depends on 21)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 21, wherein based on the improvement degrees and the comprehensive capability evaluation result of the intelligent algorithm, determining the intelligent algorithm that needs to be adjusted comprises:
if the improvement degrees of the network-level performance and/or the user-level performance are/is not changed or decreased after the optimal tuning strategy is performed, determining the comprehensive capability evaluation result of each intelligent algorithm; and
based on the comprehensive capability evaluation result of each intelligent algorithm, in combination with a corresponding evaluation criteria, determining and updating the intelligent algorithm that needs to be adjusted.

### Claim 31

An electronic device, comprising a memory, a transceiver, and a processor, wherein
the memory is configured to store a computer program; the transceiver is configured to receive and send data under a control of the processor; and the processor is configured to execute the computer program in the memory and implement following steps:
constructing the wireless communication network knowledge graph based on endogenous factors of a wireless communication network and a graph construction method, and performing deep analysis and graph cyclic tuning in combination with classified wireless communication network data;
determining a diagnostic positioning result for a wireless communication network anomaly based on the tuned wireless communication network knowledge graph, an anomaly detection algorithm and an intelligent reasoning algorithm; and
determining a plurality of tuning strategies based on a strategy generation algorithm and the diagnostic positioning result, and determining, in combination with an intelligent decision algorithm, one optimal tuning strategy from the plurality of tuning strategies by using a best result when considering an execution efficiency and an execution effect of a tuning strategy as a target, so as to perform tuning on the wireless communication network.

### Claim 32

A computer-readable storage medium, having a computer program stored therein, wherein the computer program is configured to enable a computer to execute following steps:
constructing the wireless communication network knowledge graph based on endogenous factors of a wireless communication network and a graph construction method, and performing deep analysis and graph cyclic tuning in combination with classified wireless communication network data;
determining a diagnostic positioning result for a wireless communication network anomaly based on the tuned wireless communication network knowledge graph, an anomaly detection algorithm and an intelligent reasoning algorithm; and
determining a plurality of tuning strategies based on a strategy generation algorithm and the diagnostic positioning result, and determining, in combination with an intelligent decision algorithm, one optimal tuning strategy from the plurality of tuning strategies by using a best result when considering an execution efficiency and an execution effect of a tuning strategy as a target, so as to perform tuning on the wireless communication network.

### Claim 33 (depends on 20)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 20, wherein before constructing the wireless communication network knowledge graph based on the endogenous factors of the wireless communication network and the graph construction method, and performing deep analysis and graph cyclic tuning in combination with the classified wireless communication network data, the method further comprises:
acquiring raw data of the wireless communication network, performing pre-processing on the raw data, and then performing classification storage on the preprocessed data according to different data types, wherein the classification storage of different data types is implemented based on a distributed system architecture.

### Claim 34 (depends on 19)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 19, wherein the fault type comprises a fault corresponding to different entities in the entity library and a fault corresponding to different inter-entity relationships in the inter-entity relationship library.

### Claim 35 (depends on 19)

The method for intelligently analyzing and applying the wireless communication network knowledge graph as claimed in claim 19, wherein the intelligent decision algorithm comprises one or more of: a Deep Reinforcement Learning (DRL) algorithm, a Markov Decision Process (MDP) algorithm, a biology imitation algorithm, and a statistical learning algorithm.
