Generalised Likelihood Uncertainty Estimation (GLUE) Framework Joost Delsman, Deltares, 2011 Generic Python framework to conduct GLUE analyses NOTE: Package is still under construction and undocumented Necessary: 1) model parameters, that: - hold the statistical properties of the a priori parameter space - hold the statistical properties of the a posteriori parameter space - can return a random value based on the a priori parameter space - can accept an evaluated parameter + behavioural / non-behavioural statement from the evaluator 2) a model, that: - returns a result set based on a parameter set - can be called successively to explore the parameter space - resides outside GLUE - (can even be outside python) - function or wrapper must return a dict with parameters and their values, recognised by the evaluator 3) a model evaluator, that: - evaluates the model result by a prescribed set of rules - assigns behavioural / non-behavioural 4) a framework that: - initializes a monte-carlo sequence - for each run: - get parameter values - run the model - evaluate the model - store behavioural runs