Predicting Network Enhancement Gain via Application-Level Data Trends

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional communication network optimization technologies are inefficient in detecting network capacity and coverage issues across large numbers of cells and cannot predict optimization gains related to user Quality of Experience (QoE) due to a lack of effective application-level data collection.

Innovation Solution

A method that collects network data at the application level, identifies areas with capacity or coverage problems, and predicts the gain of network enhancement operations by analyzing trends in network metrics such as download speed and network traffic, using regression analysis and weighted averages to generate optimization recommendations for cell densification and other improvements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional network optimization technologies are used to detect network capacity and coverage issues, then the detection process can be performed, but the efficiency is low and cannot effectively predict optimization gains related to user Quality of Experience

Engineering Contradiction:

Improveprediction accuracy of optimization gainsVSAvoiddetection efficiency of network issues

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent introduces an intermediary prediction model that acts as a mediator between traditional network optimization technologies and user Quality of Experience metrics. This model collects application-level data as intermediate information to bridge the gap between network infrastructure data and user-perceived performance, enabling accurate prediction of optimization gains without requiring direct measurement of all network parameters.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the traditional mechanical/manual network optimization approach with an automated prediction system using machine learning models. The system automatically collects data, performs regression analysis, and generates optimization recommendations, substituting manual network optimization processes with intelligent automated systems that can efficiently handle large-scale network data and predict outcomes.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If cell densification operations are implemented to improve network performance, then network capacity and coverage can be enhanced, but operational expenses and capital expenditures increase

Engineering Contradiction:

Improvenetwork performance and coverageVSAvoidoperational expenses and capital expenditures

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The patent applies preliminary action by performing prediction analysis before implementing cell densification operations. The system uses historical data and regression models to predict which areas will benefit most from densification, allowing network operators to plan and execute optimization operations only where they will yield the highest return on investment, thereby avoiding unnecessary capital expenditures.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent changes the parameter of decision-making from reactive to predictive by introducing predicted enhancement gain as a key parameter. Instead of implementing densification based on current network state alone, the system incorporates predicted future performance metrics, allowing operators to optimize the timing and location of infrastructure investments to maximize efficiency and minimize waste.

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If application-level data collection is implemented to predict optimization gains, then Quality of Experience prediction accuracy improves, but system complexity increases

Engineering Contradiction:

ImproveQuality of Experience prediction accuracyVSAvoiddata collection and processing system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies universality by designing a multi-functional prediction system that handles multiple types of data (network infrastructure data, application-level data, user behavior data) through a unified regression analysis framework. The same prediction model serves multiple purposes: predicting download speed gains, network traffic improvements, and overall Quality of Experience enhancements, reducing the need for separate specialized systems.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS11564112B2Communication network optimization based on predicted enhancement gain

Publication Date: 2023.01.24 META PLATFORMS INC

AI summary

In one embodiment, a computing system may receive a request for an optimization recommendation of a geographic area of interest covered by a communication network. The computing system may determine a network traffic trend associated with the geographic area of interest based on a current number of data samples that may be aggregated into a plurality of data points. The computing system may predict a value of a number of data samples for a future time associated with the geographic area of interest, based on the determined network traffic trend and the current number of data samples. The computing system may predict, based on the determined network traffic trend and the predicted value of the number of data samples at the future time, network traffic associated with the geographic area of interest at the future time, and send instructions for presenting the optimization recommendation based on the predicted network traffic.