Genetic algorithm applied to simultaneous parameter estimation in bacterial growth

Area of Science:

• Food microbiology
• Mathematical modeling
• Computational biology

Background:

• Mathematical models are crucial for understanding microbial interactions and predicting growth in food systems.
• Accurate parameter determination is essential for the predictive power of these models.
• Existing methods for parameter fitting can be time-consuming and may lack robustness.

Purpose of the Study:

• To develop and validate an inverse engineering approach using multi-objective optimization for fitting microbial growth models.
• To apply a genetic algorithm for simultaneous fitting of multiple experimental growth curves.
• To compare the efficacy of the proposed multi-objective method against conventional sequential fitting techniques.

Main Methods:

• Utilized inverse engineering combined with a multi-objective optimization procedure.
• Employed a genetic algorithm to determine optimal model parameters and construct a Pareto front.
• Applied the method to experimental data sets of co-growing lactic acid bacteria (LAB) and Listeria monocytogenes (LM).

Main Results:

• The multi-objective genetic algorithm successfully fitted multiple experimental growth curves simultaneously.
• The proposed method demonstrated superior parameter identifiability compared to conventional sequential fitting.
• The Pareto front provided a set of optimal trade-off solutions for model parameters.

Conclusions:

• The multi-objective optimization approach using a genetic algorithm offers a more robust and identifiable parameter estimation for microbial growth models.
• This method is particularly effective for modeling the simultaneous growth of interacting microbial species like LAB and LM.
• The findings suggest a significant improvement over traditional parameter fitting strategies in food safety and modeling.