Developing a model
| | | Inference | | Analysis | |
| | | ↓ | | ↓ | |
| Mechanics + | Observations | → | Model | → | Predictions |
| | ↓ | ↑ | | ↑ | ↓ |
| | → | Validation | + | Verification | ← |
Modelling is a vital part of decision-making under uncertainty. For instance, the answer to the design question "How Safe is Safe Enough?" requires unbiased predictions of hazards and impacts. This information is unavailable from many conservative engineering models and code equations. To deal with this situation, this toolbox provides an open library of probabilistic models that are intended to simulate physical phenomena. All uncertainty, both irreducible aleatory uncertainty and reducible epistemic uncertainty in the models are characterized by random variables. As a result, the models can be analyzed by reliability methods to compute event probabilities. In fact, Rt is tailored for reliability and optimization analyses with multiple probabilistic models.
Utilizing a model
| Random variable realizations | → | | | |
| Design variable realizations | → | Model | → | Unique realization |
| Upstream model responses | → | | | |
|