การปรับปรุงกระบวนการวิวัฒนาการในวิธีกำหนดการเชิงพันธุกรรมเพื่อสร้างคำตอบที่มีความทนทานสำหรับปัญหาการนำร่องหุ่นยนต์ / รุ่งโรจน์ นพสุวรรณชัย = Improving evolutionary process in the genetic programming for generating a robust solution to the robot navigation problem / Roongroj Nopusuwanchai
This research presents improvement of the Genetic Programming learning method for generating a robust solution to the robot navigation problem. The solution is in the form of a computer program that controls the robot to go to a target while avoiding obstacles within a particular environment. The robustness of a solution is defined as an ability to perform successfully in the environment that is different from the one that it was trained. The robustness of the solution is increased by improving the evolutionary process during the evolution of solutions by learning from multiple environments and by increasing the difference among those training environments. The robustness of a solution before and after improving evolutionary process is compared. The result shows that increasing the number of environments and increasing the difference among environments cause the solution to be more robust. The analysis shows that the robustness of a solution depends on the acquired ‘experience’ during the evolution and the ratio of experience reusing in an unknown environment.
