A novel surrogate-based optimisation framework for pharmaceutical process systems.

Area of Science:

• Chemical Engineering
• Process Systems Engineering
• Computational Chemistry

Background:

• Pharmaceutical manufacturing relies on complex process models for optimization.
• High computational demands of these models necessitate efficient alternative approaches.
• Surrogate-based optimization offers a practical solution for complex system modeling.

Purpose of the Study:

• To present a unified framework integrating multiple software tools for surrogate-based optimization.
• To address challenges in optimizing complex real-world process models.
• To improve key metrics like yield, purity, and sustainability in Active Pharmaceutical Ingredient (API) manufacturing.

Main Methods:

• Development of a unified system for surrogate-based optimization.
• Application of multi-objective optimization techniques.
• Utilizing Pareto fronts for visualizing trade-offs between competing objectives.

Main Results:

• Surrogate models effectively approximate complex process behaviors.
• Single-objective optimization improved Yield by 1.72% and Process Mass Intensity by 7.27%.
• Multi-objective optimization enhanced Yield by 3.63% while maintaining high purity.

Conclusions:

• The presented framework provides a practical approach for robust optimization of API manufacturing flowsheets.
• Surrogate-based optimization is an efficient strategy for improving pharmaceutical process performance.
• Multi-objective optimization effectively balances competing performance metrics in drug manufacturing.