Neighboring crossover to improve GA-based Q-learning method for multi-legged robot control

ABSTRACT

In this paper, we propose a crossover method to improve a GA-based Q-learning method for controlling multi-legged robots. As a GA-based Q-learning method, we employ a method called "Q-learning with Dynamic Structuring of Exploration Space Based on Genetic Algorithm (QDSEGA)". We propose a crossover for QDSEGA, and a method to reward a robot in Q-learning in order to follow a moving target. Simulation results clearly show the effectiveness of the proposed methods.

REFERENCES

• 1: Ito, K. and Matsuno, F. A study of reinforcement learning for the robot with many degrees of freedom -Acquisition of locomotion patterns for multi legged robot-, In Proc. of IEEE Int'l Conf. on Robotics and Automation, pp. 392--397, 2002.
• 2: Richard S. Sutton, Reinforcement Learning, Kluwer Academic Publishers, Norwell, MA, 1992
• 3: Christopher J. C. H. Watkins , Peter Dayan, Technical Note: $\backslash cal\; Q$-Learning, Machine Learning, v.8 n.3-4, p.279-292, May 1992 [doi>10.1007/BF00992698]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Knowledge acquisition

General Terms:
Algorithms, Performance

Keywords:
Q-learning, Q-table restructuring by genetic algorithm, multi-legged robot control