Li Shoutao and Gordon Lee
Evolutionary Algorithms for Learning Primitive Fuzzy Behaviors and Behavior Coordination in MultiObjective Optimization Problems
560 - 565
2012
6
6
International Journal of Mechanical and Materials Engineering
https://publications.waset.org/pdf/1813
https://publications.waset.org/vol/66
World Academy of Science, Engineering and Technology
Evolutionary robotics is concerned with the design of
intelligent systems with lifelike properties by means of simulated
evolution. Approaches in evolutionary robotics can be categorized
according to the control structures that represent the behavior and the
parameters of the controller that undergo adaptation. The basic idea
is to automatically synthesize behaviors that enable the robot to
perform useful tasks in complex environments. The evolutionary
algorithm searches through the space of parameterized controllers
that map sensory perceptions to control actions, thus realizing a
specific robotic behavior. Further, the evolutionary algorithm
maintains and improves a population of candidate behaviors by
means of selection, recombination and mutation. A fitness function
evaluates the performance of the resulting behavior according to the
robots task or mission. In this paper, the focus is in the use of
genetic algorithms to solve a multiobjective optimization problem
representing robot behaviors; in particular, the ACompander Law is
employed in selecting the weight of each objective during the
optimization process. Results using an adaptive fitness function show
that this approach can efficiently react to complex tasks under
variable environments.
Open Science Index 66, 2012