Travelling salesman problem

Result type
software
Description

Active Hopfield neural network dynamics algorithms implementing gradient descent with gradient energy function of the network and simulated annealing of graph node permutations in the form of dynamically linked libraries. The traveling salesman problem is an NP-hard discrete optimization problem, mathematically expressing and generalizing the problem of finding the shortest possible path passing through all vertices of a ranked graph. In practice, such a problem is usually solved only approximately by heuristic algorithms such as genetic algorithms, simulated annealing or continuous Hopfield networks. This achieves (at the cost of giving up the claim of finding an optimal solution) practical times. It can be used, for example, to optimize the order of visits to different facilities for the purpose of revising them with respect to the reviser's transport costs.

Solution of the travelling salesman problem using optimization model (black) and simulated annealing (blue)

Downloads:

The software has been developed with a financial support of the Technology Agency of the Czech Republic under the research project No. TK04020003.

Keywords
Traveling Salesman Problem; Hopfield Neural Network; Simulated Annealing
Main benefits

Savings on transport costs.