Traveling Salesman Problem Solver
This NeuronDotNet application uses Kohonen SOM to find an approximate solution to traveling salesman problem.
Download LinksThe objective of traveling salesman problem (TSP) is to find a shortest path which starts from any city, goes through all cities in the sales map and returns back to the starting city. Formally, we have to find a shortest Hamiltonian cycle in an undirected weighted complete graph containing every city as a node, the distance between two cities being the weight of the edge incident on corresponding nodes. This problem is known to be NP-Complete.
Here, we attempt to find an approximate solution to TSP using Kohonen Self-organizing maps. The training data consists of X and Y co-ordinates of all cities in a randomly generated sales map. A Kohonen SOM with a 'Circular Layer' (a layer with unit height and circular rows) when trained with this training set, organizes itself such that it represents an approximate solution to the TSP.
Zedgraph, an open source library is used to plot the graphs.
A screenshot of the application is shown below.
