James Silverton -

I take it that Solver has not solved the "Travelling salesman" problem
while I was not looking? <

As far as I know, the main difficulty with solving the Traveling Salesman
Problem (TSP) is that for an n-city problem there are (n-1)! possible tours.
Thus, a 17-city TSP has 21 trillion possible routes. So, the problem can
take a long time to solve.

It's easy to set up the TSP in a spreadsheet, and it's in most management
science and operations research textbooks. Those textbooks usually include
Premium Solver for Education, which has an "alldifferent" constraint and an
Evolutionaly Solver that are not in the standard Solver shipped with Excel.
The TSP can be set up and solved using standard Solver, but the alldifferent
constraint simplifies the spreadsheet formulation.

On the www.solver.com web site, if you search locally for TSP, you'll see
several examples, including one that says "The classic Traveling Salesman
Problem (TSP) is easy to solve using the "alldifferent" constraint supported
by all Solver Engines in the Solver Platform SDK."

- Mike

http://www.MikeMiddleton.com




"James Silverton" wrote in message
...
Mike wrote on Tue, 10 Mar 2009 11:32:32 -0700:

In addition to the examples in the SolvSamp workbook
(available on your hard drive if you have installed Solver
from the CD), there are more examples available on the Solver web site:

Distribution Examples:
http://www.solver.com/discenter.htm

Scheduling Examples:
http://www.solver.com/schcenter.htm

Larger problems may exceed the capacity of standard Solver
(200 adjustable cells), but Frontline Systems provide more
advanced versions of Solver.

I take it that Solver has not solved the "Travelling salesman" problem
while I was not looking? Isn't there a Fields medal going?
--

James Silverton
Potomac, Maryland

Email, with obvious alterations: not.jim.silverton.at.verizon.not