Neural Network Traffic Modeling for Adaptive Capacity Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for modeling network traffic capacity are inadequate for complex, large-scale global networks, as they rely on inaccurate assumptions of self-similarity and long-range dependency, struggle with variable data characteristics, and are not adaptable to changing network conditions.
Innovation Solution
The use of artificial neural networks to decompose and model network traffic into categories, allowing for intelligent and adaptive prediction of network capacity based on input attributes, which can adjust to changes over time and location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional mathematical models based on fractional Brownian motion are used to model network traffic, then the modeling process is simple and algorithmic, but the accuracy deteriorates when network traffic characteristics change over time and location
Solution Approach 1:
The patent replaces conventional mathematical models (fractional Brownian motion) with artificial neural networks. The neural network uses learning algorithms to adaptively model network traffic patterns, substituting the rigid mathematical framework with a flexible computational system that can capture complex, non-stationary traffic characteristics without requiring explicit mathematical formulations.
Solution Approach 2:
The patent introduces dynamic adaptability into the modeling process by training neural networks on historical traffic data. The model automatically adjusts its parameters and structure based on learned patterns, enabling it to adapt to changing network traffic characteristics over time and location, thereby maintaining high prediction accuracy in dynamic environments.
2Ease of operation
If conventional mathematical equations are used for network capacity determination, then the calculation process is deterministic and straightforward, but the system cannot adapt to variable data characteristics and changing network conditions
Solution Approach 1:
The patent implements feedback mechanisms through the training process, where the neural network continuously learns from historical network traffic data. The model uses feedback from actual traffic patterns to adjust its predictions, enabling it to adapt to variable data characteristics and changing network conditions while maintaining operational straightforwardness through automated learning processes.
Solution Approach 2:
The neural network performs self-service by automatically learning and adapting to network traffic patterns without requiring manual reconfiguration. The system self-adjusts its parameters based on training data, providing adaptability to changing conditions while maintaining ease of operation through automated, self-contained learning processes.
3Stability of the object's composition
If network traffic is modeled using fixed mathematical assumptions, then the modeling approach is simple and consistent, but it fails to accurately represent variable and evolving traffic patterns in complex global networks
Solution Approach 1:
The patent transitions from static mathematical assumptions to dynamic neural network models that evolve with training data. The neural network maintains consistency in its learning framework while adapting its internal parameters to accurately represent variable traffic patterns, achieving both modeling consistency and representation accuracy in complex global networks.
Solution Approach 2:
The patent changes the fundamental parameters of the modeling approach by replacing fixed mathematical assumptions with learnable neural network parameters. This allows the model to maintain structural consistency while adapting its parameters to accurately capture evolving traffic patterns, improving representation accuracy without sacrificing modeling coherence.
Data Source
AI summary
A method and system are provided for modeling network traffic in which an artificial neural network architecture is utilized in order to intelligently and adaptively model the capacity of a network. Initially, the network traffic is decomposed into a plurality of categories, such as individual users, application usage or common usage groups. Inputs to the artificial neural network are then defined such that a respective combination of inputs permits prediction of bandwidth capacity needs for that input condition. Outputs of the artificial neural network are representative of the network traffic associated with the respective inputs. For example, a plurality of bandwidth profiles associated with respective categories may be defined. An artificial neural network may then be constructed and trained with those bandwidth profiles and then utilized to relate predict future bandwidth needs for the network.