Title :
A Harmony Theory network solution to the map-coloring problem
Author :
Tambouratzis, T.
Author_Institution :
Inst. of Inf. & Telecommun., NCRPS, Attiki, Greece
Abstract :
A parallel implementation of the map-coloring problem is presented. The problem consists of assigning a color to each region of a map so that: (a) the least number of colors is employed, (b) each region is attributed a unique color, and (c) no two adjacent regions are assigned identical colors. This optimization problem is formulated as an incremental iterative constraint-propagation task and is encoded in a Harmony Theory network. This choice was made since Harmony Theory has proved to be a particularly efficient tool for solving constraint-propagation tasks in parallel. Consequently, the potential of Harmony Theory as a parallel tool for solving optimization problems is investigated.
Keywords :
graph colouring; graph theory; mathematics computing; neural nets; optimisation; Harmony Theory network; incremental iterative constraint-propagation task; map-coloring problem; optimization problem; parallel tool; Approximation algorithms; Artificial intelligence; Computational modeling; Constraint optimization; Constraint theory; Explosions; Informatics; Mathematics; Parallel processing; Simulated annealing;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.716880