Genetic Algorithms with Random Coordinates for Route Planning on a 3D Terrain

In this paper, an approach using genetic algorithms (GA) is proposed to find a feasible and optimum route on a 3D Geographical Information System (GIS) platform. The approach is based on generating random coordinates on the region instead of using discrete grids. Specialized crossover and mutation operators are defined accordingly. The feasibility criterion is to prevent the route from passing through the water. The two criteria to be optimized are the distance of the route from starting point to the destination point and the amount of climb ups/downs through the route. Our method is tested on NASA´s World Wind application, which is an open source 3D virtual globe developed in Java.