KPI Optimization via Influential Parameter Change Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in optimizing the state of a target, such as a manufacturing plant, due to the difficulty in selecting appropriate parameters from a large number of variables, which can lead to complex relationships and minimal parameter changes required to achieve optimal Key Performance Indicators (KPIs).
Innovation Solution
An information processing apparatus and method that calculates prediction values for each explanatory variable based on a model representing the state of a target, and determines the change quantity of the objective variable with changes in the explanatory variables, allowing for the identification of influential parameters and optimization of KPIs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all parameters are changed to optimize the KPI, then the optimal KPI can be obtained, but the complexity of the optimization process increases and some parameters cannot be changed depending on situations
Solution Approach 1:
The patent segments the optimization process into two distinct stages: first identifying candidate parameters through correlation analysis with the KPI, then optimizing only those selected parameters. This segmentation resolves the contradiction by filtering out parameters that cannot be changed or are less influential, thereby reducing optimization complexity while maintaining effectiveness.
Solution Approach 2:
The patent applies local quality by treating different parameters differently based on their correlation with the KPI. High-correlation parameters are selected for optimization while low-correlation parameters are excluded. This selective approach optimizes the process by focusing computational resources on the most influential parameters rather than treating all parameters uniformly.
2Ease of operation
If the number of parameters to be selected is narrowed down, then the optimization process becomes easier, but it is difficult to select an appropriate number of parameters from 1000 or more parameters due to complicated relationships
Solution Approach 1:
The patent uses correlation analysis as a feedback mechanism to evaluate the relationship between each parameter and the KPI. By calculating correlation coefficients, the system provides quantitative feedback about parameter importance, enabling automated selection of the most influential parameters without requiring manual assessment of complicated parameter relationships.
Solution Approach 2:
The patent replaces manual parameter selection with an automated computational system that uses correlation analysis and statistical methods. This substitution transforms the difficult manual task of evaluating 1000+ parameters into an automated process that objectively identifies influential parameters based on their mathematical relationship with the KPI.
3Productivity
If minimal parameters are changed to obtain optimal KPI, then the optimization efficiency improves, but it is difficult to identify which parameters are most influential
Solution Approach 1:
The patent introduces correlation coefficients as an intermediary metric that quantifies the relationship between parameters and the KPI. This intermediary provides objective information about parameter influence, enabling the system to identify and select the most influential parameters for optimization without losing critical information about parameter importance.
Data Source
AI summary
An information processing apparatus according to the present invention includes: a prediction value calculation unit configured to, based on a model calculating an objective variable representing a state of a target by using a plurality of explanatory variables, for each of the explanatory variables, calculate a prediction value of the explanatory variable that changes based on a value of the explanatory variable at a predetermined moment; and a change quantity calculation unit configured to, for each of the explanatory variables, calculate a quantity of change of the objective variable with change of the explanatory variable by using the prediction value of the explanatory variable and the model.