Using passive neural networks to solve TSP

Numeric experiments on random, relatively large travelling salesman problems presented in this paper show that passive neural networks can be used as an efficient, dynamic optimization tool for combinatorial programming.