Cognitive Autonomous Network Controller for Optimal Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-organizing networks (SON) face limitations in adapting to rapidly changing environments due to their rule-based nature and the complexity of maintaining and upgrading numerous rules, which hinders optimal configuration determination in cognitive autonomous networks (CANs).
Innovation Solution
Implementing a controller in CANs that determines optimal configurations based on proposed configurations from multiple network automation functions (NAFs) with differentiated interests, by accounting for individual interests and calculating an interest-weighted optimal value for configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If rule-based self-organizing networks (SON) are used for network automation, then network configuration can be managed systematically, but the system cannot adapt rapidly to changing environments and becomes complex to maintain and upgrade
Solution Approach 1:
The patent implements a self-learning mechanism where the system automatically learns optimal configurations through reinforcement learning algorithms. The network automation function continuously observes network states, receives rewards or penalties based on performance, and updates its own configuration policies without human intervention, enabling rapid adaptation while reducing maintenance complexity
Solution Approach 2:
The system dynamically adjusts configuration parameters based on learned patterns and changing network conditions. Instead of fixed rule-based configurations, the system modifies parameters in real-time through continuous learning, allowing rapid adaptation to environmental changes while the underlying learning model remains maintainable
2Productivity
If multiple network automation functions (NAFs) with differentiated interests are deployed, then network functionality and coverage are enhanced, but determining optimal configurations becomes more complex due to conflicting interests
Solution Approach 1:
The patent merges multiple NAFs into a unified reinforcement learning framework where all functions share a common learning agent or coordinated set of agents. This integration allows the system to handle multiple differentiated interests through a single optimization process that considers all functions' goals, reducing the complexity of determining optimal configurations while maintaining enhanced network functionality
Solution Approach 2:
The system implements a universal configuration determination mechanism that serves multiple NAFs with different interests. The reinforcement learning framework is designed to handle diverse function types and interests through a unified approach, allowing the same system to optimize configurations for multiple functions simultaneously without requiring separate complex determination processes for each
Data Source
AI summary
Systems, methods, apparatuses, and computer program products for determining and/or applying optimal configurations in cognitive autonomous networks (CANs) are provided.