Parameter identification for chemical reaction systems using sparsity enforcing regularization: a case study for the...

Area of Science:

• Chemical kinetics
• Computational chemistry
• Systems biology

Background:

• Complex chemical reactions are modeled using nonlinear ordinary differential equations.
• Parameter identification from experimental data is challenging due to data limitations and uncertainties, leading to non-unique and unstable solutions.
• Minimizing data mismatch alone is insufficient for reliable parameter identification.

Purpose of the Study:

• To develop a regularization method that promotes sparsity for parameter identification in chemical reaction models.
• To eliminate unidentifiable model parameters, reducing model complexity while maintaining consistency with experimental data.
• To enhance the reliability and interpretability of chemical kinetic models.

Main Methods:

• A sparsity-promoting regularization approach is proposed.
• The adjoint state technique is utilized for efficient gradient computation with respect to parameters and initial conditions.
• The method is demonstrated on the chlorite-iodide reaction system.

Main Results:

• The proposed method effectively identifies essential model parameters by eliminating those with low sensitivity to the data.
• The resulting reduced models are more manageable and interpretable.
• The approach ensures model consistency with experimental observations.

Conclusions:

• Sparsity-promoting regularization offers a robust solution for parameter identification in complex chemical systems.
• This method simplifies complex models by focusing on the core reaction mechanisms.
• The technique enhances the reliability of parameter estimation and model interpretation.