An approach for solving dynamic TSPs using neural networks
Kazuhiko Shinozawa, Tadasu Uchiyama, Katsunori Shimohara · 1991
An approach for solving the D-TSP (dynamic traveling salesman problem) is discussed. The difference between a TSP and a D-TSP is that, in the latter, traveling costs between cities change according to the time of day. A neural network was applied to a D-TSP. This D-TSP was given a two-dimensional representation. With several maps, minimum cost tours were found by computer simulation using a Hopfield-Tank-type neural network.>