Bayesian Multiobjective Optimisation With Mixed Analytical and Black-Box Functions: Application to Tissue...

Area of Science:

• Biomedical Engineering
• Regenerative Medicine
• Computational Biology

Background:

• Tissue engineering aims to restore biological function using scaffolds and bioreactors.
• Optimizing neotissue growth in bone tissue engineering faces conflicting goals: maximizing growth and minimizing cost.
• Current methods often rely on less data-efficient algorithms like genetic algorithms.

Purpose of the Study:

• To develop and evaluate a novel Bayesian multiobjective optimization method.
• To optimize design variables for maximizing neotissue growth while minimizing operating costs in bone tissue engineering.
• To address optimization problems combining analytical and black-box (simulation-based) objective functions.

Main Methods:

• Extended Bayesian multiobjective optimization for mixed analytical and black-box objectives.
• Simulation of a partial differential equation system to model neotissue growth.
• Comparison of the proposed method's data efficiency against genetic algorithms.

Main Results:

• The developed method effectively determines the trade-off between neotissue growth and operating cost.
• Demonstrated superior data efficiency compared to genetic algorithms on a bone tissue engineering case study.
• Validated performance on standard test functions, showing improved data efficiency.

Conclusions:

• The novel Bayesian multiobjective optimization approach offers a more data-efficient solution for complex engineering problems.
• This method is applicable to real-world scenarios involving simulation-based models and quantifiable objectives like cost.
• Provides a robust framework for balancing competing objectives in regenerative medicine and tissue engineering.