AI Control Model Selection for Fast Effect Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face difficulties in easily verifying the effect of AI control introduction to a system, requiring skilled engineers to manually adjust parameters and generate complex models, increasing cost and time.
Innovation Solution
An information providing apparatus and method that automatically generates and evaluates multiple models based on operational data, determining the best models for prediction and control, and generates an AI control introduction effect report.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual model generation and parameter adjustment is performed by skilled engineers, then model accuracy and reliability are improved, but device complexity and time consumption increase
Solution Approach 1:
The system performs automatic model generation, evaluation, and selection without requiring manual intervention by skilled engineers. The automated model generation unit creates multiple candidate models, the evaluation unit assesses them using operational data, and the selection unit automatically identifies the optimal model, enabling the system to serve itself rather than relying on external expert intervention
Solution Approach 2:
The system automatically adjusts and optimizes model parameters through computational evaluation of multiple candidate models. By generating models with varying parameters and systematically evaluating their performance against operational data, the system identifies optimal parameter configurations without manual adjustment, reducing both time consumption and dependency on expert knowledge
2Reliability
If manual model generation and parameter adjustment is performed by skilled engineers, then model quality is improved, but device complexity increases
Solution Approach 1:
The system replaces the mechanical process of manual model generation and parameter adjustment with an automated computational system. Instead of engineers manually creating and tuning models, the system uses algorithmic model generation units and automated evaluation mechanisms to produce and assess models, substituting human expert processes with systematic computational procedures
Solution Approach 2:
The system generates multiple candidate models as copies with varying parameters and structures. By creating these replicated model versions and systematically evaluating them, the system identifies the optimal model without requiring manual intervention, reducing process complexity while maintaining model quality through automated comparison and selection
3Ease of operation
If automated model generation is implemented, then ease of operation is improved, but manufacturing precision may deteriorate
Solution Approach 1:
The system generates multiple candidate models beyond what would be minimally required, creating an excess of model options. By generating more models than strictly necessary and systematically evaluating them, the system ensures that the optimal model is identified through comprehensive comparison, maintaining precision while improving ease of operation through automation
Solution Approach 2:
The system incorporates an evaluation unit that assesses each generated model using operational data from the target system. This feedback mechanism allows the system to automatically determine model quality by comparing predictions against actual operational outcomes, ensuring manufacturing precision is maintained while enabling ease of operation through automated evaluation and selection
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A server device (10) generates N simulation models on the basis of operational data collected from a system, determines a simulation model in which a simulation model evaluation value for evaluating the simulation model is a maximum value on the basis of the operational data, generates M AI control models on the basis of the operational data and the determined simulation model, and determines the AI control model in which an AI control model evaluation value for evaluating the AI control model is a maximum value on the basis of the operational data.