Implementation of Morris method for sensitivity analysis

[zepholus]

@debpal Should we add the Morris method to the SensitivityAnalyzer module?
Perhaps adding a method key inside simulation_by_sample_parameters would be enough to let the user choose between Morris or Sobol (for now).
This would allow users to perform faster sensitivity analyses when needed, since Sobol and Morris require very different numbers of samples.

• Sobol generally requires a large number of samples, typically N × (2k + 2), where k is the number of parameters and N is the base sample size (often 1000).
• Morris, on the other hand, is much more efficient. It usually needs only r × (k + 1) samples, where r is the number of trajectories (often 50), resulting in far fewer total runs.

Adding the method option would allow users to balance accuracy (Sobol) and speed (Morris) depending on their needs.

Checklist

• [X ] I have searched for existing feature requests to avoid duplicates.