Graph coloring using fuzzy controlled neural networks
(EN)
Dalianis, P
(EN)
Tzafestas, S
(EN)
Kitsios, Y
(EN)
Graph coloring is an important example of a class of combinatorial optimization problems, which are characterized by their large number of interacting degrees of freedom. Neural Networks (NN) and especially a variation of the Hopfield one constitutes a very popular approach to the solution of problems in this class. The algorithm presented here is based on the idea of improving the convergence performance of such a NN, with the use of a simple fuzzy controller. The algorithm is explained and applied to two graphs of different size. The simulation results are presented and described in detail.
(EN)