10/25 Do any websites (ala Mapquest or Etak) approximate The Travelling
Salesman Problem? I went to visit 15 condos in San Diego today, and
it was a bitch using Yahoo+Thomas guide to map out a semi-ideal
route. I don't need an optimum path, just a first order approximation.
I don't think I'll ever need more than 20 addresses at once either...
I'd pay $.99 every time I need this feature...
\- the emacs mapquest interface can do that.
do M-x mapquest-metropolis --psb
\_ I knew emacs was cool. I didn't know how cool until now.
\_ Is this the real psb? Because that didn't work for me.
\- you need the Los Alamos Simulated Annealing emacs lisp
package. --psb
\- that was not the real psb. ok tnx. -psb
\- no, that was me. the immediately above is not me. --psb
\_ You mean it was !psb, right psb?
\_ What is the "Travelling Salesman Problem"?
\_ No cookie.
\_ find out the route which traverses all cities with the minimum
distance.
\_ Just use the thomas guide. It's at 1200dpi. mapquest is at
whatever resolution your monitor is at.
\_ you're a moron
\_ My point is that there is WAY more detail on a single
thomas guide page than on 30 yahoo maps put together.
\_ Okay, get a Real map (Thomas or AAA) big enough
to fit all the points. Then put a dot sticker at
each address and use your highly evolved
mamalian intellect to see patterns. Connect the
dots. Do you have a beter solution?
dots. It's not perfect, but do you have a solution?
\_ You're still a moron. The OP wants to know how to get
from point A -> B -> ... -> Z in an optimal manner.
\_ Hey person with Etak friend... can you ask them if they have
anything like that. Or is there any API where a 3rd party could
program that feature?
\_ I can't wait until <DEAD>maps.google.com<DEAD>... google would support it!
\_ Even your first order solution will be near useless because of the
ugly realities of traffic patterns, highway congestion, etc. I'll
get 30 miles down a freeway at 3am faster than you'll get 10 miles
down a dirt road at noon. Just use some common sense.
\- the algorithms allow for weights ... of coure knowing what
weights to plug in in a little tricky. i am not sure the
algorithms i know of can deal with "asymmetric" weights ...
usually on a metric the distance from d(x,y) = d(y,x) which
clearly isnt true with traffic flows if we care about time.
although some LA trunks seem to be equally congested in both
directions. --psb
\_ When you're taking city streets within a 5 mile range, it
isn't that big of a difference. Or at least not in my case.
\_ isn't there a way to plot multiple locations at once? just do that
and do a rough aproximation yourself. the above poster is right...
small distance details will be less important than type of road,
traffic, etc. But just plotting them all on one map should be
enough for you to use common sense to plan a route.
\_ This would work. Which service supports this (yahoo doesn't)? |