> > Very unlikely that A* beats CH in general. > It certainly doesn't. > > > A* is just so much more flexible. For example, you can add a new vehicle type > without any significant penalty in terms of preprocessing. Or avoid traffic > jams that change frequently. And the flexibility can result in significant > savings on fuel and time.
I grant A* that it does not require any special preprocessing. I also grant that CH is not the answer to all questions. Avoiding traffic jams is a much more delicate topic, because the naive method of adding a penalty and recomputing a new path usually fails. Also, you want to have some guarantee on alternative path quality and also on runtime. A* does not guarantee anything. > Furthermore, as computers get cheaper and devices get faster, the difference > in speed between A* and CH is becoming insignificant. Plain wrong. A* does not scale. > All these tradeoffs are firmly in favour of A* for most applications. Also wrong. Especially in applications you want to have small search spaces and also compute many (sub-)paths.
_______________________________________________ Routing mailing list Routing@openstreetmap.org http://lists.openstreetmap.org/listinfo/routing
