Handling of optimisation results is unclear

[katielukow-AE]

Feature description

It would be great if the examples were more clear on how to handle the resulting output from the optimisation class. For example, how to access the PyBaMM solution with the identified parameter set.

Motivation

To provide insight on how to handle the resulting identified parameters outside of built-in plots

Possible implementation

No response

Additional context

No response

[NicolaCourtier]

Hi @katielukow-AE, thanks for opening this issue! The code has changed quite a lot since, so I wanted to check what (if any) improvement you would still like to see.

The typical workflow now involves creating a PyBaMM model, simulator and optimisation problem. Running an optimiser or sampler returns a result for the given problem, which includes the identified best_inputs.

The result can be used for further investigation, for example:

1. Reproduce the optimised solution using simulator.solve(inputs=result.best_inputs)

2. Reproduce the cost value (and sensitivities) using problem.evaluate(inputs=result.best_inputs)

3. Continue with PyBaMM using parameter_values.update(result.best_inputs) then e.g. create a new simulation with pybamm.Simulation(model, parameter_values=parameter_values)

Do these examples resolve the issue?