An optimized differential evolution algorithm for constitutive model fitting of arteries.

Area of Science:

• Biomedical Engineering
• Materials Science
• Computational Mechanics

Background:

• Arterial mechanical properties are crucial for cardiovascular health.
• Biaxial testing and constitutive models are standard for assessment.
• Current fitting algorithms may not yield optimal results.

Purpose of the Study:

• To introduce an optimized fitting routine for improved constitutive model accuracy.
• To evaluate the efficacy of Differential Evolution (DE) and novel loss functions.
• To enhance the description of human superficial femoral artery mechanical behavior.

Main Methods:

• Acquired experimental stress-stretch data from 16 human superficial femoral arteries via biaxial testing.
• Employed the Differential Evolution (DE) metaheuristic algorithm with Minkowski and Chebyshev distance loss functions.
• Utilized grid search for hyperparameter tuning to optimize algorithm settings.

Main Results:

• The DE algorithm consistently outperformed traditional fitting algorithms across all samples.
• Minkowski and Chebyshev distance-based loss functions yielded better fits than square error.
• Achieved superior fitting quality for constitutive models of arterial tissue.

Conclusions:

• The proposed DE-based fitting routine significantly enhances constitutive model accuracy for arterial mechanics.
• Novel loss functions improve the fit, enabling simpler constitutive relations with fewer parameters.
• This approach increases the efficiency of computational implementations in biomechanics.