Machine Learning Network Configuration Automation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Network engineers face a time-consuming and labor-intensive process in determining optimal configuration parameters for communication networks to improve performance, relying on empirical methods rather than automated solutions.

Innovation Solution

A machine learning-based apparatus that predicts performance indicators for communication networks by analyzing configuration parameters and network characteristics, identifying dominant features, and recommending influential parameter values to enhance network performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If network engineers use empirical methods to determine configuration parameters, then they can improve network performance based on experience, but the process becomes time-consuming and labor-intensive

Engineering Contradiction:

Improvenetwork performanceVSAvoidconfiguration determination time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system enables self-service by allowing the network to automatically determine optimal configuration parameters through machine learning models. The apparatus autonomously analyzes performance indicators, identifies dominant features, and recommends configuration changes without requiring manual intervention from network engineers, thus resolving the contradiction between maintaining reliable performance improvement and reducing time consumption.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical/manual process of empirical configuration determination with an automated machine learning-based system. The machine learning model substitutes the human engineer's empirical analysis with algorithmic processing, automatically predicting performance indicators and identifying influential configuration parameters, thereby eliminating the time-consuming manual trial-and-error process while maintaining performance optimization.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Manufacturing precision

If network engineers manually adjust configuration parameters, then they can optimize network settings, but the process requires significant manual effort and expertise

Engineering Contradiction:

Improveconfiguration optimization precisionVSAvoidconfiguration process ease

Core Design Contradiction:

Manufacturing precisionVSEase of operation

Solution Approach 1:

The system performs self-service by automatically analyzing performance indicators and determining optimal configuration parameters without human intervention. The machine learning model autonomously identifies dominant features and generates configuration recommendations, eliminating the need for manual effort and specialized expertise while maintaining high precision in configuration optimization.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The machine learning model acts as an intermediary between network performance data and configuration parameters. It processes performance indicators, identifies influential features, and translates this analysis into specific configuration recommendations, thereby automating the complex optimization process and making it accessible without requiring deep manual expertise from engineers.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Measurement precision

If a machine learning model uses multiple features for prediction, then prediction accuracy improves, but the complexity of analyzing dominant features increases

Engineering Contradiction:

Improveperformance indicator prediction accuracyVSAvoidfeature analysis complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system extracts and focuses on dominant features from the multiple input features by analyzing feature importance scores. The machine learning model identifies and isolates the most influential configuration parameters that have the greatest impact on performance indicators, extracting only the critical subset of features needed for effective prediction and recommendation, thereby managing complexity while maintaining accuracy.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system applies local quality by treating different features with different levels of importance. Instead of uniformly processing all features, the machine learning model identifies dominant features with higher local quality (greater influence on predictions) and focuses computational resources on these key features, thereby improving prediction accuracy while reducing the effective complexity of feature analysis.

Inventive Principle:

Principle #3Local quality

Data Source

PatentEP3849231B1Configuration of a communication network

Publication Date: 2022.07.06 NOKIA SOLUTIONS & NETWORKS OY

Figure 1~2
Figure 3
Figure 4

AI summary

An apparatus determines a configuration for a communication network. The apparatus implements a machine learning model to predict a value of a performance indicator of the communication network. The machine learning model is configured to predict the value of the performance indicator based on a plurality of features. The features include data related to configuration parameters of the communication network. The configuration parameters can be modified through a management apparatus of the communication network. The apparatus further exploits the machine learning model to determine one or more recommended values for one or more configuration parameters. The recommended values for the configuration parameters improve the predicted value of the performance indicator.