Graph Neural Network Traffic Steering for Energy-Aware Handovers

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing communication networks face challenges in optimizing network efficiency and power consumption due to exponential network traffic and complex task processing, particularly in advanced networks like 5G and 6G, leading to suboptimal resource allocation, increased handover failures, and inefficient spectrum utilization.

Innovation Solution

A data-driven traffic steering approach using gradient boosted decision tree models and graph neural networks to optimize network traffic across cells, balancing user equipment quality of service and energy consumption by predicting optimal handovers based on spatial dependencies and user equipment-centric metrics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If traditional network traffic management is used, then network traffic can be handled, but power consumption increases exponentially and resource allocation becomes suboptimal

Engineering Contradiction:

Improvepower consumptionVSAvoidnetwork efficiency

Core Design Contradiction:

Loss of energyVSProductivity

Solution Approach 1:

The patent implements dynamic traffic steering that adapts network resource allocation in real-time based on changing traffic patterns and network conditions. The system continuously monitors and adjusts traffic routing decisions, transitioning from static to dynamic management to optimize the balance between energy consumption and network efficiency

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes key operational parameters including traffic steering decisions, handover thresholds, and resource allocation metrics based on learned patterns from graph neural networks. By dynamically adjusting these parameters rather than using fixed configurations, the system reduces power consumption while maintaining network productivity

Inventive Principle:

Principle #35Parameter changes

2Productivity

If complex task processing is implemented to handle exponential network traffic, then network capacity increases, but power consumption and processing requirements increase significantly

Engineering Contradiction:

Improvenetwork capacityVSAvoidpower consumption

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The system applies partial action by implementing traffic steering only where and when needed rather than processing all network traffic uniformly. Graph neural networks identify specific traffic flows and network segments that benefit from enhanced processing, avoiding unnecessary complex operations on routine traffic and thereby reducing overall power consumption while maintaining network capacity

Inventive Principle:

Principle #16Partial or excessive action

3Productivity

If frequent handovers are performed to optimize network traffic distribution, then resource allocation improves, but handover failures increase and network stability decreases

Engineering Contradiction:

Improveresource allocationVSAvoidhandover failure rate

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system implements feedback mechanisms where graph neural networks continuously learn from handover outcomes and network performance metrics. By analyzing past handover successes and failures, the system adjusts future handover decisions to improve success rates while maintaining optimized resource allocation. This closed-loop approach reduces handover failures compared to purely proactive handover strategies

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary evaluation of handover candidates using graph neural network predictions before executing actual handovers. By pre-assessing the likelihood of successful handovers and preparing target cells in advance, the system reduces handover failures and improves reliability while maintaining the benefits of optimized resource distribution

Inventive Principle:

Principle #10Preliminary action

4Productivity

If advanced machine learning models are deployed for traffic steering, then network optimization improves, but system complexity and computational requirements increase

Engineering Contradiction:

Improvenetwork optimizationVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the network into graph-based representations where nodes and edges capture essential network relationships. This segmentation allows graph neural networks to process network state information in a structured, scalable manner, improving optimization capabilities while managing system complexity through modular graph-based architecture rather than monolithic processing

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250310786A1Facilitating facilitating dynamic network traffic steering using graph neural networks in advanced communication networks

Publication Date: 2025.10.02 DELL PROD LP

AI summary

Facilitating dynamic network traffic steering using graph neural networks in advanced communication networks is provided. A method includes determining respective results of application of a utility function to all combinations of potential handovers of the specified user equipment from a source cell to respective target cells of a group of target cells. A communication network can include the source cell and the group of target cells. The operations can also include, based on the respective results of the application of the utility function, implementing a network traffic steering process that moves connectivity of the user equipment from the source cell to a single target cell of the group of target cells.