Real Time Pathfinding with Genetic Algorithm
Alex Fernandes da Veiga Machado, Ulysses O. Santos, H Vale, Rubens Goncalvez, Tiago Araújo Neves, Luiz Satoru Ochi, Esteban Clua · 2011
This paper presents a method to optimize the process of finding paths using a model based on genetic algorithms and A* for real time systems, such as video games, virtual reality environments. The proposed solution uses obstacle pattern detection based on online training system that is generally used in systems with real time requirements and in dynamic environments. The architecture, named Real Time Path finding with Genetic Algorithm (RTP-GA), uses a Genetic Algorithm in order to create an agent adapted to the environment that is able to optimize the search for paths even in the presence of obstacles. In specific cases, the RTP-GA architecture presents a complexity that is better than A* algorithm.