Related Experiment Videos
Optimal experiment design for nonlinear models subject to large prior uncertainties.
The American Journal of Physiology
|September 1, 1987
Summary
This study introduces a robust experiment design method that accounts for unknown parameter uncertainty by optimizing expected Fisher information. This approach simplifies robust design, making it comparable to classical D-optimal design for applications like multiexponential model timing.
Area of Science:
- Statistics
- Experimental Design
- Mathematical Modeling
Background:
- Classical experiment design often relies on unknown parameter values.
- A priori parameter uncertainty is a significant challenge in experimental planning.
Purpose of the Study:
- To develop a robust experiment design methodology that incorporates prior parameter uncertainty.
- To optimize the mathematical expectation of a functional of the Fisher information matrix.
Main Methods:
- Utilizing a stochastic approximation algorithm for optimization.
- Integrating prior parameter statistics into the design process.
- Applying the methodology to determine optimal measurement times for multiexponential models.
Main Results:
- The proposed method effectively accounts for a priori parameter uncertainty.
- Robust experiment design is achieved with comparable simplicity to D-optimal design.
- The methodology provides a framework for optimizing measurement schedules in complex models.
Conclusions:
- The new approach offers a practical solution for robust experiment design under parameter uncertainty.
- This method enhances the reliability of experimental outcomes, particularly for multiexponential models.
- It bridges the gap between classical optimal design and the need for uncertainty handling.