Abstract: |
The Ant Colony Metaheuristic was originally proposed for tackling optimization problems. More recent research has suggested that it can be applied for automatic generation of programs. By allowing the artificial ants to visit functions and terminals nodes, they become able to build pheromone trails that represent computer programs for optimizing a fitness domain-specific function. In this paper a novel approach is addressed using a grid architecture as a more suitable discrete world to be explored by the ants. The resulting system was applied to automatically produce programs to solve Boolean functions. |