It is also important to note that, unless you mod the DB to correct this, the method Garfunkel describes using INF+LOG-S is almost strictly superior to any method using LVH+LOG.
The reason is that INF+LOG-S costs 0.2 BP for a 10-ton unit providing 100 GSP, so you get 500 GSP per build point. LVH+LOG costs 2.48 BP for a 62-ton unit providing 500 GSP, so you get just under 200 GSP per build point. Thus, infantry logistics are 250% more efficient per build point, and just under 25% more tonnage-efficient as well. Prior to 1.12 this was 'balanced' by the fact that only LVH logistics could resupply other formations automatically (the micro for infantry logistics was not feasible at scale), but now with the unit replacement mechanic the LVH logistics are utterly outclassed and obsolete.
The only reason infantry logistics are not strictly superior is that you are required to put some amount of supply units in your frontline formations, perhaps 5-10% of the tonnage to ensure that your units remain supplied between two construction ticks (which is when the unit replacement step occurs). This is assuming 5-day construction increments, so if you use a lower increment as many do then the requirement is absolutely minimal. Otherwise, LVH logistics do benefit from being resistant to enemy fire but this is not worth 250% greater build cost given the cost of necessary supplies to actually mount an invasion.
Given this, I mod my DB so that the 50-ton LOG component gives 1000 GSP instead of 500 and make it LVH-only, which gives both infantry and LVH logistics different pros and cons thus restoring an interesting decision. Infantry logistics remain BP-efficient by about 25%, but LVH are more tonnage-efficient by about 40%. In this case a mix of frontline LOG-S elements, HQ LVH+LOG, and LOG-S supply dumps which can be transported during or after an invasion are all useful components of a complete logistics pipeline.
