first, a couple points about your post ...
"it is unlikely that critical fastener will deform as much as the low loaded fasteners" ... the idea behind load proportional to distance is that bolts will strain proportional to distance, therefore more highly loaded bolts are expected to strain more.
i think you can do something about load redistribution, about fasteners becoming plastic and so more flexible compared to elastic bolts. i think it's somewhat reasonable to imagine that the bolt group won't Fail untill all the fasteners are working at their ultimate allowable. i think it'll become a messy calc ...
apply your elastic dist'n untill one bolt reaches it's allowable, say 80% of the required load
delete the critical fastener from the load calc, essentially limiting it to it's allowable, then re-do fastener loading to balance the new pattern about the old CG, apply maybe 5% load or untill the next fastener reaches it's allowable,
and so on !
it might be easier to say each fastener can carry it's ultimate allowable, acting in generalised directions (each fastener reacts Pult in direction "alpha", which varies with each fastener, yes?), so that the fastener forces balance P and P*d (where d is the off-set in your design, to say the bolt group elastic CG) so you'd get to ultimte allowable "P" for the joint.
this approach is probably optimistic, you might consider the effect if the critical bolt fails.
clear as mud ?