(calibration_minimisation)=
# Using minimisation techniques

The theory behind this method may seem very basic as it consists mainly in a point estimation of a 
valid configuration that could be performed without accounting for underlying uncertainty. This view 
is overly simplistic, since numerical optimisation problems are not simple, even for 
single-objective optimisation due, for instance, to:

- the regularity of the cost function, the presence of possible local minima, and the computational cost involved;

- the dimensionality of the input space and the potentially complex constraints on the inputs.

These challenges can be addressed through an appropriate problem formulation and the use of a suitable 
optimisation algorithm. However, these choices are not always straightforward, especially when 
calibration is confronted with identifiability issues. For further methodological aspects on these issues, 
see [](#optimisation_module).