Title :
Distribution of asymptotically stable states in Hopfield network for TSP
Author :
Matsuda, Satoshi
Author_Institution :
Comput. & Commun. Res. Center, Tokyo Electr. Power Co. Inc., Japan
Abstract :
Taking traveling salesman problem (TSP) as an example, it is theoretically shown how asymptotically stable states are distributed in the network state hypercube, and that their solution qualities are quantitatively characterized by their locations in hypercube. These characterization makes dynamics of network clear. Moreover, optimal network, which most effectively gets optimal solutions, is also theoretically characterized
Keywords :
Hopfield neural nets; Hopfield network; asymptotically stable state distribution; dynamics; hypercube; network state; optimal network; traveling salesman problem; Asymptotic stability; Constraint optimization; Hypercubes; Intelligent networks; Neurons; Space exploration; Stability analysis; Traveling salesman problems;
Conference_Titel :
Neural Networks, 1996., IEEE International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-3210-5
DOI :
10.1109/ICNN.1996.548949
