Surrogate Model Optimization Using Feasible Parameter Sets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optimization methods for manufacturing or controlling technical systems are resource-intensive and often generate parameter sets that cannot be evaluated or do not satisfy system constraints due to insufficient information about the computerized model, leading to inefficient and ineffective parameter settings.
Innovation Solution
A computer-implemented method that uses a surrogate model generated by a regression method based on feasible parameter sets to determine an optimized parameter set, incorporating a feasibility identifier to ensure the model accounts for system constraints, thereby accelerating the optimization process and improving the generation of suitable parameter sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a typical optimizer is used without sufficient information about the computerized model, then the optimization process can be run, but the optimizer generates parameters that cannot be evaluated or do not satisfy system constraints, resulting in useless parameter sets
Solution Approach 1:
The patent applies preliminary action by pre-processing the sample parameter sets through a feasibility identifier before generating the surrogate model. This preliminary classification separates feasible from infeasible parameter sets, ensuring the surrogate model is trained only on valid data. This prevents the optimizer from generating useless parameter sets and improves both efficiency and reliability of the optimization process.
2Measurement precision
If computerized simulation methods such as FEA or CFD are used to evaluate parameter values, then accurate evaluation is achieved, but the process becomes time and resource consuming
Solution Approach 1:
The patent creates a surrogate model as a simplified copy of the complex computerized simulation model. This surrogate model replicates the essential evaluation functionality but with significantly reduced computational requirements. The optimization process uses this copy instead of the original complex simulation, achieving accurate parameter evaluation while dramatically reducing time and resource consumption.
Solution Approach 2:
The patent transforms the complex simulation model into a surrogate model by changing the parameters and structure of the evaluation system. The surrogate model uses simplified mathematical relationships and pre-computed data structures instead of running full FEA or CFD simulations, maintaining evaluation accuracy while reducing computational complexity and execution time.
3Quantity of substance
If the surrogate model is generated using all parameter sets including erroneous and non-feasible ones, then more data is available for training, but the model accuracy decreases and generates invalid optimization results
Solution Approach 1:
The patent applies local quality by differentiating the quality of training data through the feasibility identifier. Instead of treating all parameter sets uniformly, the system identifies and selectively uses only the high-quality feasible parameter sets for surrogate model training. This localized quality filtering ensures the surrogate model learns from valid examples, improving model accuracy and preventing the propagation of errors in optimization results.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This invention relates to a computer-implemented method for manufacturing or controlling a technical system, comprising the method steps: (a) inputting (S1) a sample of parameter sets suitable for manufacturing or controlling the technical system, with a feasibility identifier assigned to each of the parameter sets, wherein the feasibility identifier marks each parameter set either as technically feasible or technically non-feasible or erroneous in terms of manufacturing or controlling the technical system, (b) generating (S2) a computerized surrogate model for the technical system based on the respective parameter sets of the sample, which are marked as technically feasible, by means of a regression method, (c) determining (S3) an optimized parameter set based on the surrogate model by means of a computerized optimization method, and (d) outputting (S4) the optimized parameter set for manufacturing or controlling the technical system.