AI Graph Search for Communication Network Failure Investigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search methods for failure investigation ranges in communication networks are inefficient, particularly in large-scale networks, due to complex conditional branches, reliance on operator skill, and high computational demands of graph search algorithms.
Innovation Solution
A search device and method that generates a graph of connected devices within an initial investigation range, using AI-estimated failure locations, and extends this range based on inference from a search model trained on past failure results, incorporating neighboring devices as needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a rule-based workflow is used to search for failure investigation range, then comprehensive investigation is possible, but the system complexity and difficulty of creation increase significantly
Solution Approach 1:
The patent replaces the mechanical rule-based workflow system with a machine learning model that automatically learns failure propagation patterns from historical data. The ML model substitutes complex conditional rules with data-driven predictions, maintaining comprehensive investigation capability while eliminating workflow complexity.
Solution Approach 2:
The system enables self-service by allowing the machine learning model to automatically determine investigation ranges without human intervention in workflow design. The model self-adjusts and optimizes investigation strategies based on learned patterns, eliminating the need for manual workflow creation and maintenance.
2Reliability
If graph search algorithm is used to search for failure investigation range, then systematic search is possible, but calculation amount and time increase enormously in large-scale networks
Solution Approach 1:
The patent applies preliminary action by pre-training the machine learning model on historical failure data before actual failure investigation. The model learns propagation patterns in advance, enabling it to quickly predict investigation ranges during actual failures without performing exhaustive graph searches at runtime.
Solution Approach 2:
Instead of performing complete exhaustive graph searches, the system uses the ML model to predict only the necessary investigation range based on learned patterns. This partial action approach focuses computational resources on the most relevant devices, avoiding unnecessary searches in large-scale networks.
3Reliability
If depth-first search is performed to explore all possible failure paths, then complete investigation is achieved, but calculation amount becomes enormous
Solution Approach 1:
The patent extracts the essential failure propagation patterns from historical data using the machine learning model. By extracting these patterns beforehand, the system avoids re-processing complete failure paths during actual investigations, significantly reducing computational energy while maintaining complete investigation coverage.
Data Source
AI summary
A search device 1 for searching for an investigation range of a failure occurring in a communication network includes a generation unit 16 configured to generate a graph in which a plurality of devices within a certain investigation range including a suspected failure location obtained by artificial intelligence (AI) are connected on the basis of a connection configuration of devices constituting a communication network, anda search unit 17 configured to input the graph to a search model capable of searching for an investigation range on the basis of past failure results and to cause the search model to infer whether to extend the investigation range of the graph, and the generation unit 16 adds a neighboring device adjacent to a device in the graph to the graph in a case in which the investigation range of the graph needs to be extended.