A branch and bound solution to the travelling salesman problem. To find the best path, the program traverses a tree that it creates as it goes. It uses a lower bound cost algorithm to prune paths who couldn't possibly be lower than the current best path. To initialize the best cost, a greedy solution is found.
Cities and their distances are stored in a file that is selected at runtime. A sample file, tsp/cities.txt, contains the coordinates from 15 cities in Luxembourg and demonstrates the file format. The format is the standard TSPLIB format and is as follows:
NAME: <name>
COMMENT: <comment>
TYPE: TSP
DIMENSION: <n>
EDGE_WIEGHT_TYPE: EUC_2D
NODE_COORD_SECTION
<1> <x> <y>
<2> <x> <y>
...
<n> <x> <y>
EOF
The headers and EOF are all optional but the loader will use them if specified.
The program can be run with the cities file as an argument or a dialog will ask for the file. The command line arguments are:
java -jar TSP.jar cities.tsp.txt