Considering that a magnetic coil acts on a iron object in such a fashion as to always try to maximise its inductance, then there must be a load of design optimisations for the coil geometry that depend upon the velocity of the projectile? Ie you want to be be able to build the maximum coil current before the projectile reaches the max inductance location, and then ideally, immediately reduce the magnetic flux to zero to avoid creating a drag force on the projectile as it exits the coil. Hence, so sort of iron magnetic flux steering might make an improvement?
The problem with recovering some of the firing energy, using an H bridge type driver is that the voltage available to reduce the magnetic flux is limited to that of the storage medium. Using a simple on/off switch (of a suitably high voltage rating) means the flux can be destroyed by a much higher voltage that that used to generate that flux (there is no penalty (in terms of barrel velocity) with turning on coils early, compared to a significant one with turning them off late!
In all cases, you need to maximise the amp turns for any given inductance value