A Harmony Theory network solution to the map-coloring problem

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.