Global Versus Local Constructive Function Approximation for On-Line Reinforcement Learning.

Peter Vamplew Robert Ollington Global Versus Local Constructive Function Approximation for On-Line Reinforcement Learning. 113-122 2005 conf/ausai/2005 Australian Conference on Artificial Intelligence http://dx.doi.org/10.1007/11589990_14 db/conf/ausai/ausai2005.html#VamplewO05