On the Asymptotic Behaviour of a Constant Stepsize Temporal-Difference Learning Algorithm.

Vladislav Tadic On the Asymptotic Behaviour of a Constant Stepsize Temporal-Difference Learning Algorithm. 126-137 http://link.springer.de/link/service/series/0558/bibs/1572/15720126.htm 1999 conf/eurocolt/1999 EuroCOLT db/conf/eurocolt/eurocolt1999.html#Tadic99