This paper presents a practical application of the evolution strategy (ES) to the evolution and optimization of Braitenberg vehicles. Braitenberg vehicles are a special class of autonomous agnets. Autonomous agents are embodied systems that behave in the real world without any human contorl. One major goal of research on autonomous agents is to study intelligence as the result of a system environment interaction, rather than understanding intelligence on a computational levle. Braitenberg vehicles are controlled by a number of parameters, which are mostly determined by hand in a trial and error process. This paper shows that a simple ES evolves Braitenberg vehicles very efficiently. Other research has used genetic algorithms (GAs) for very similar tasks. A comparison of both approaches shows that the ES approach is much more efficient. Since autonomous agents are very important in the field of new AI, this research field should spend more attnetion to evolution strategies.
展开▼
