The NEOS Server offers Concorde for the solution of
Traveling Salesman Problems.
Concorde was written by
David Applegate,
Robert E. Bixby,
Vašek Chvátal, and
William J. Cook.
Source code, binaries, and documentation are available from the
Concorde homepage
This solver was implemented by
Hans Mittelmann
and executes at
Using the NEOS Server for Concorde
The user must submit a symmetric TSP problem in either the simple 2d coordinate form (first line is just the number of cities!)
#cities
x_0 y_0
x_1 y_1
.
.
.
x_n1 y_n1
or in
TSPLIB format.
The user can currently choose between applying the exact algorithm and the
LinKernighan heuristic (especially for large instances).
Concorde can be called with the authors'
QSopt LP solver
or
CPLEX.
This small
benchmark gives you an
impression of its performance with different LP solvers.
If Concorde terminates prematurely it may have run out of time or memory.
Either the fixed random seed 99 can be used or a variable one.
Users who submit via web submission (not email or XMLRPC) can further
choose to receive a PDF file plotting the optimal resp. final tour.
Enter the complete path to the file with the xylist (distances measured in the Euclidean(L2) norm)
Concorde data(xylist file, L2 norm):
Or, enter the complete path to the file with the xylist (distances measured in the Manhattan(L1) norm)
Concorde data(xylist file, L1 norm):
Or, enter the complete path to the symmetric TSPLIB file
Concorde data(TSPLIB format file):
Choose the algorithm (cqs=QSopt, con=CPLEX [default], lk=LinKernighan)
Algorithm:
Choose the random seed (fixed=99 [default], variable=random)
Random seed:
PDF file of the optimal tour? (no [default], cp=yes, pf=w/o cities)
PDF plot:
Comments:
Dry run: generate job XML instead of submitting it to NEOS
Short Priority: submit to higher priority queue with maximum CPU time of 5
minutes
Please do not click the 'Submit to NEOS' button more than once.
