Adaptive Search Region Control for Faster Process Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search methods for optimizing processing conditions in devices like chemical reaction devices and semiconductor processing devices are inefficient, often taking a long time to reach the optimum solution due to the method's reliance on simply changing the search region based on the difference between actual and target values.
Innovation Solution
A search device and method that utilize a prediction model to efficiently search for the optimum input parameters by dynamically updating the search region based on the improvement rate of actual measurement values, allowing for a more focused and rapid convergence to the target output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the search region is simply changed based on the difference between actual measurement value and target value, then the search method is easy to implement, but it takes a long time to reach the optimum solution
Solution Approach 1:
The patent applies dynamics by making the search region adaptive and dynamically adjustable. Instead of using a fixed or simply changed search region, the system dynamically determines the search region based on the improvement rate calculated from actual measurement values and prediction values. This allows the search process to adapt its scope and focus based on real-time feedback, resolving the contradiction between ease of implementation and search efficiency.
Solution Approach 2:
The patent implements feedback mechanisms by calculating the improvement rate from the difference between actual measurement values and prediction values, then using this feedback to determine the next search region. This closed-loop feedback system enables the search to learn from previous results and adjust accordingly, significantly reducing search time while maintaining ease of operation through automated feedback processing.
2Manufacturing precision
If the search region is frequently adjusted to reduce the difference between actual measurement value and target value, then the accuracy of reaching the optimum solution is improved, but the search speed decreases
Solution Approach 1:
The patent applies partial action by strategically selecting when and how to adjust the search region based on the improvement rate. Instead of frequently adjusting the search region for every minor improvement, the system adjusts only when the improvement rate indicates a significant opportunity for optimization. This partial adjustment approach maintains accuracy while preserving search speed by avoiding unnecessary adjustments.
Solution Approach 2:
The patent utilizes parameter changes by modifying the search region parameters based on the calculated improvement rate. When the improvement rate indicates promising directions, the search region parameters are adjusted to focus on those areas; when the improvement rate is low, the search region remains stable. This dynamic parameter adjustment resolves the contradiction by changing parameters only when it benefits both accuracy and speed.
3Device complexity
If the search method uses a simple region change approach, then the device complexity is low, but the search efficiency is insufficient
Solution Approach 1:
The patent implements self-service by enabling the search system to automatically determine and adjust its own search region based on calculated improvement rates from measurement data. The system serves itself by autonomously making decisions about search region adjustments without requiring complex external control mechanisms, thereby maintaining low device complexity while significantly improving search efficiency through intelligent self-adjustment.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided are a search device, an operation method of a search device, an operation program of a search device, and flow reaction equipment, which can improve a search speed. A prediction data set generation unit generates a prediction data set composed of a plurality of prediction data in which an explanatory variable for which a value of a response variable is unknown and a prediction value of the response variable are associated with each other by using a known data set. A first actual measurement value acquisition unit acquires an actual measurement value of the response variable included in the prediction data in which the prediction value is closest to a target value. An improvement rate calculation unit calculates an improvement rate representing a difference between a known value of the response variable closest to the target value and the actual measurement value. A known data set update unit adds the actual measurement value and a value of the explanatory variable corresponding to the actual measurement value to the known data set in a case in which the improvement rate is equal to or higher than a target improvement rate. A second actual measurement value acquisition unit acquires an actual measurement value of the response variable for a value of the explanatory variable included in the prediction data, which is not used for acquiring the actual measurement value by the first actual measurement value acquisition unit, in a case in which the improvement rate is lower than the target improvement rate.