| On 2008-06-10 at 14:00:00 (Brussels Time) |
Abstract
A heuristic approach to automated planning is proposed. A planning problem consists in finding a sequence of actions, or instances of operators, which transforms an initial state to a desired goal state. Many algorithms and representations was proposed to deal with. In particular planning graph is a compact structure to encodes a planning problem and most of its constraints (mutex relations). We are investigating the application of "Ant Colony Optimization (ACO)-inspired" algorithms to extract valid plans from a planning graph. Ant Colony Optimization is a robust heuristic that has been succesfully applied to many combinatorial optimization problems. Some considerations and suggestions will be presented.
Keywords
Ant Colony Optimization, AI planning
