Add local optimisation methods and penalty functions to MDMC and derivatives of FoM

**Is your feature request related to a problem? Please describe.**
In most cases, users should have a set of force fields, such as OPLS, but need to refit them for some reason. In most cases, they will not want to deviate far from the original parameterisation to avoid overfitting, especially since MDMC appears to be designed to fit a few experimental results at most. We should allow the user to choose to run a local optimisation rather than the much more expensive global optimisation procedure currently used in MDMC. Additionally, we may want to add some penalty-type function to avoid overfitting, similarly to what is done in forcebalence; see https://pubs.acs.org/doi/10.1021/jz500737m.

Also consider adding a way to calculate the derivative of the FoM with respect to the force field parameters using Eq. 2 from https://pubs.acs.org/doi/10.1021/jz500737m.

**Describe the solution you'd like**
Add local optimisation routines to MDMC. Add penalty-type functions to the figure of merit.

Start by locating MDMC’s current global optimisation routine and the figure-of-merit implementation; the issue names no files or tests. Review how force-field parameters and experimental results flow through those entry points, then define completion around selectable local optimisation, penalty functions, and a force-field-parameter FoM derivative.

Written by the indexing model from the issue text.