This seems not fully worked out. Assuming total travel is constrained, selectively limiting travel to particular areas concentrates travel in the remaining areas, which increases the average number of contacts between travelers, which violates your “reducing exposure to other people is best” rule of thumb.
By your model, some reductions in travel ought to have a negative effect. Is that a fair corollary?