Adaptive Parameter Search Space for Faster Optimization Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In industries with high interaction effects among operating parameters, defining an effective parameter search range is challenging, leading to either convergence to non-optimal local solutions or the need for numerous trials, which increases costs and delays product delivery.
Innovation Solution
A system and method for parameter optimization with an adaptive search space, utilizing a data acquisition unit, parameter space transformer, and search range definer to dynamically adjust the parameter search range based on executed values, reducing the interaction effect and optimizing the target parameter efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the parameter search range is defined too narrow, then the optimization converges faster, but it may converge to a non-optimal local solution
Solution Approach 1:
The patent applies dynamics by making the parameter search range adaptive rather than static. The system dynamically adjusts the search range based on the current optimization state and historical data, allowing it to expand when needed to avoid local optima and contract when converged to accelerate optimization. This is implemented through the adaptive search range determination module that modifies search boundaries during the optimization process.
Solution Approach 2:
The patent changes the parameters of the search space itself. By transforming the original parameter space into a new coordinate system and adaptively adjusting the search range boundaries based on optimization progress, the system effectively changes the search parameters dynamically. This allows the search to cover appropriate regions at different optimization stages, balancing exploration and exploitation.
2Reliability
If the parameter search range is defined too wide, then the optimization can find the global optimal solution, but the number of trials must be quite large
Solution Approach 1:
The system dynamically adjusts the search range width based on optimization progress. Initially, a wider search range is used to ensure global optimality, but as optimization progresses and the system gains information about the parameter landscape, the search range is adaptively contracted to reduce the number of required trials while maintaining convergence to the optimal solution.
Solution Approach 2:
The patent performs preliminary actions by transforming the parameter space and determining an adaptive search range before the main optimization process. This preliminary space transformation and range determination prepares the optimization system to search more efficiently by pre-identifying promising regions and establishing appropriate search boundaries based on initial data analysis.
3Reliability
If the number of trials is increased to find the optimal solution, then the solution quality improves, but it affects the delivery time and causes cost waste
Solution Approach 1:
The patent changes the parameter representation through space transformation. By transforming to a new coordinate system where the relationship between parameters and target values is more clearly revealed, the system can identify optimal solutions with fewer trials. This parameter transformation enables more efficient exploration of the parameter space, improving solution quality without proportionally increasing trial numbers.
Solution Approach 2:
The system uses feedback from executed trials to adaptively adjust the search range. As trials are executed and results obtained, this feedback information is used to refine and contract the search range, eliminating regions that are unlikely to contain the optimal solution. This feedback-driven adaptation reduces the total number of trials needed while maintaining high solution quality.
Data Source
AI summary
A system and a method for parameter optimization with adaptive search space and a user interface using the same are provided. The system includes a data acquisition unit, an adaptive adjustment unit and an optimization search unit. The data acquisition unit obtains a set of executed values of several operating parameters and a target parameter. The adaptive adjustment unit includes a parameter space transformer and a search range definer. The parameter space transformer performs a space transformation on a parameter space of the operating parameters according to the executed values. The search range definer defines a parameter search range in a transformed parameter space based on the sets of the executed values. The optimization search unit takes the parameter search range as a limiting condition and takes optimizing the target parameter as a target to search for a set of recommended values of the operating parameters.