Optimization on unbounded solution space using dynamic genetic algorithms

The problem of applying genetic algorithm (GA) to solve an optimization problem over an unbounded solution space is addressed. We propose to first transform the possible range of each parameter in the chromosome to a finite range with a nonlinear mapping such that the search on unbounded solution space becomes a search for high precision solution in a finite range. Modifications on the GA have been found necessary after such nonlinear mapping. As a result, a new GA with dynamic mutation range that facilitates coarse-refine search has been developed