Merging incomparable linearizations

If you have L1=[5,3,1,8,0] (chunk feerates: 5, 4, 0) and L2=[0,8,1,3,5] (chunk feerates: 4, 3) then bestPi chooses C=L2 \cap [5] to start with, ending up as [5,8,3,1,0]. Whereas calculating and comparing C_1 and C_2 would produce [8,5,3,1,0].

But post-processing probably fixes that for most simple examples at least?

(Even if post-processing didn’t fix it; I think so long as merging produces something at least as good as both, that’s fine; quicker/cheaper is more important)