Probabilistic Cost Function Inference for System Parameter Tuning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining optimized system parameters for complex technical systems is challenging due to the lack of clear, automated methods for defining cost functions, which are often subjective and require expert intuition and manual adjustment, making the process costly and time-consuming.
Innovation Solution
A computer-implemented method that automatically determines optimized system parameters using a cost function by generating probability functions based on historical system parameters and output values, combining them to maximize overall probability, and optimizing system parameters without human experimentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If expert knowledge and manual adjustment are used to determine cost functions, then the cost function can be adapted to specific requirements, but the process becomes costly and time-consuming
Solution Approach 1:
The system performs self-service by automatically generating cost functions through probabilistic reasoning over historical data and rules, eliminating the need for manual expert iteration. The automated system determines optimized parameters and adapts to new requirements without human intervention, resolving the contradiction between adaptability and time consumption.
Solution Approach 2:
The system performs preliminary action by pre-processing historical system parameters and output values into structured rules and probability functions. This preparation work is done in advance, enabling rapid automatic determination of cost functions when new parameter optimization is needed, thus reducing the time and cost of subsequent adjustments.
2Productivity
If automated methods are used to determine system parameters, then the process becomes efficient and reproducible, but the ability to handle abstract expert knowledge and subjective criteria is limited
Solution Approach 1:
The system uses rules as an intermediary between abstract expert knowledge and automated optimization. Experts formulate knowledge in the form of rules with input and output variables, which serve as a bridge that allows automated probabilistic reasoning to handle subjective criteria while maintaining efficiency and reproducibility.
Solution Approach 2:
The system changes parameters by transforming abstract expert knowledge into concrete probabilistic functions with specific parameters. The probability functions have quantifiable parameters that can be automatically optimized, enabling the system to handle abstract knowledge through concrete mathematical representations.
3Extent of automation
If a clear and reliable cost function definition is provided for automatic evaluation, then automated optimization is enabled, but the requirement for precise definition of all relevant influences increases complexity
Solution Approach 1:
The system segments the cost function definition into multiple independent rules, each handling a specific aspect of system optimization. This segmentation allows automated evaluation of each rule separately while maintaining overall system complexity at manageable levels, enabling automation without requiring a single complex monolithic cost function.
Data Source
AI summary
A method for determining optimized system parameters of a technical system using a cost function. The cost function is provided for determining optimized system parameters of the technical system. The technical system has system parameter-adjustable components. When the system parameters are set, the technical system generates component output values. The method includes:providing historical system parameters and component output values,providing technical system rules based on the system parameters and their output values,determining a function space corresponding to a function set wherein the cost function lies,generating probability functions from the historical system parameters and corresponding output values using a rule, wherein each probability function indicates the probability which satisfies the rule by any cost function,combining probability functions to determine the cost function by maximizing overall probability of all rules,optimizing the system parameters given the cost function, andoutputting the optimized system parameters to adjust the components.