Project Tideline

Overview

Project Tideline is the laboratory's computational backbone: a transport-simulation front end paired with trained surrogate models, so a design sweep that used to cost a week of cluster time finishes in the time it takes to have lunch.

The surrogates are not a replacement for the simulation. They are a way to explore a design space quickly and then spend the expensive compute only where it matters, with every surrogate prediction carrying an uncertainty that is checked against the full calculation.

This paragraph is placeholder text. Replace it with the real scope of the software programme and who it is intended to serve.

Approach

Simulation front end

Placeholder text describing the interface, which solvers are supported, and how inputs are specified and versioned.

Surrogate models

Placeholder text describing the model family, the training data, and how uncertainty is estimated and validated.

Reproducibility tooling

Placeholder text describing environment pinning, provenance capture, and how a published figure is tied back to the run that produced it.