Far East Journal of Experimental and Theoretical Artificial Intelligence
-->
Abstract: The application of a Hybrid Genetic Hill-climbing
Algorithm (HGA) for n-region
4-coloring map problems is proposed. To effectively reduce the magnitude of the
search space by ~23 times, for large n, where a genetic grouping representation
that does not result in any loss of generality is proposed. The efficacy of the
HGA embodying the proposed genetic grouping representation and the proposed
objective function is
depicted by comparing its performance against the established standard
Genetic Algorithm, Hill-climbing, an artificial neural network optimization
algorithm and HGA without the embodiment of the proposed genetic grouping
representation for several n-region 4-coloring maps. The HGA with the
proposed embodiment of genetic grouping representation is shown to be the only
algorithm that is able to obtain an optimal solution for large maps, for example
where The proposed HGA is also shown to
require the shortest computation time to yield an optimal solution for up to
1200-region 4-coloring maps.
Keywords and phrases: hybrid genetic algorithm, 4-coloring map, genetic grouping representation, multi-point search and optimization.
