perfusio documentation¶ Getting Started Getting Started Installation Quick-start Reproducing the paper figures Running the CLI Citation Theory Model Theory Overview Mechanistic Model Stepwise-GP Residual Layer References Bayesian Experimental Design Objective Functions Acquisition Functions Pareto Optimality Evaluation Metrics Relative Root-Mean-Square Error (rRMSE) Prediction Interval Coverage Sharpness CRPS Multi-Objective Metrics Tutorials T01 — Simulate a Training Experiment What the script does Running Key code Design space Expected output Next step T02 — Fit the Hybrid GP Model What the script does Running Key code GP architecture Expected metrics (held-out run 023) Next step T03 — Bayesian Experimental Design (Single Objective) What the script does Running Key code Available acquisition functions Acquisition parameters Next step T04 — Multi-Objective BED and Pareto Exploration What the script does Running Key code Pareto dominance criterion Multi-objective acquisitions Next step T05 — Transfer Learning Across Cell Lines What the script does Running Key code Architecture Clone registry When to use transfer learning Next step T06 — Online Digital Twin (Filesystem Connector) What the script does Running Key code Digital twin control loop Online retraining Audit trail Next step T07 — Real ambr®250 via OPC UA What the script does Running Key code OPC UA connector features Node configuration Security considerations Design Guides Choosing an Acquisition Function Decision Tree Quick Reference Recommended Defaults Comparison with DataHowLab / Datahow High-Level Summary Design Philosophy When to use perfusio When to use DataHowLab (or similar) Regulatory Considerations Is perfusio GxP-validated? Using perfusio in a Research / Development Context Path to GMP Deployment Audit Trail Data Integrity Contact API Reference API Reference Project Changelog Limitations Contributing to perfusio