Re-define the steering force to allow for an easier definition of the biasing constant

[DSilva27]

If we define the steering energy as $E = k \text{RMSD}(X, X_{ref})$, then the force has norm $k n^{-1/2}$, with $n$ being the number of atoms. This makes the norm of the force independent of the configuration. Since the MD term is pretty stable, this allows for a more systematic definition of the steering force.

• Change $\text{RMSD}^2$ to $\text{RMSD}$
• Increase the number of steps performed when initializing a projector so that the simulations are more stable.
• Change the user input to be a constant of proportionality between the MD energy and the bias energy, i.e., "make the bias energy be 1/10 of the MD energy".
• Write a function that estimates the scaling factor given a PDB, and introduce it to the ensemble optimization command.

[DSilva27]

Before merging, create a branch with name preprint and mention in the README that the current version is not compatible with what is available there. Explain how to change the configs to match current version.