The trouble with P=NP problems is simple: and it's the same
  problem with ALL of them: Too much is left to assumption and not
  defined. The problems aren't intrinsicly difficult. No problem
  is. The problem is the CONSTRAINTS are assumed. One assumption
  is the granularity. How big? How small? What is being measured?
  If the travelling salesman is a giant, well he can visit all of
  them at once with his foot. If the travelling salesman was a
  beam of light, he could visit all nearly simultaneously, so long
  as he was up high enough to capture everybody at the same time.
  Be specific about the constraints and THEN we have something to
  work with. How wide is the road? (bandwidth). etc FURTHEST TO
  NEAREST, (DOWN THE DIAGANAL)