Title :
A new technique for reinforcement learning for control
Author :
Armstrong, William W. ; Li, Darwin
Author_Institution :
Dept. of Comput. Sci., Alberta Univ., Edmonton, Alta., Canada
Abstract :
In this work reinforcement learning was successfully applied to several simulated control problems including pendulum swing-up, pole balancing and a difficult challenge that can roughly be described as balancing a basketball on one´s finger. Compared to the first two tasks, the ball-balancing task is much harder. Control is achieved by using a piecewise linear approximant of the Q-function learned by an adaptive logic network (ALN) which solves Bellman´s equation in a high-dimensional space
Keywords :
learning (artificial intelligence); learning systems; nonlinear control systems; ALN; Bellman equation; Q-function; adaptive logic network; ball-balancing task; control; high-dimensional space; pendulum swing-up; piecewise linear approximant; pole balancing; reinforcement learning; Adaptive control; Adaptive systems; Computational modeling; Equations; Fingers; Learning; Logic; Piecewise linear approximation; Piecewise linear techniques; Programmable control;
Conference_Titel :
Systems, Man, and Cybernetics, 1998. 1998 IEEE International Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
0-7803-4778-1
DOI :
10.1109/ICSMC.1998.728123
