Aims: To develop a computationally efficient minimal cardiac model that reproduces the electrophysiological behavior of the complex ToR-ORd model.
Methods: We implemented weighted multi-objective optimization using differential evolution to adjust the four-equation Bueno-Orovio minimal model to match the 44-equation ToR-ORd model. Sensitivity analysis identified key influential parameters (theta_w, u_o, theta_o for APD90; theta_w, u_u for maximum dV/dt; u_o for peak voltage). The optimization targeted APD90 restitution curves, maximum dV/dt, and action potential morphology across cycle lengths from 300-900 ms.
Results: Optimization completed in 46,289 seconds with objective function value of 0.114. Key parameter changes included u_o (-88.76 to -89.78 mV), u_u (124.67 to 137.95 mV), and theta_w (-61.36 to -84.18 mV). The optimized model reproduced the ToR-ORd restitution curve with APD90 values from 160.75 ms (BCL=300 ms) to 261.29 ms (BCL=900 ms), with deviations below 4% at cycle lengths above 500 ms. Maximum dV/dt values matched within 3% for BCLs above 450 ms.
Conclusion: Multi-objective optimization successfully balanced electrophysiological characteristics in a minimal model that preserves essential properties of the ToR-ORd model while reducing computational complexity by a factor of 11. This enables accurate cardiac simulations with enhanced computational efficiency for large-scale applications.