mlpy.mdp.discrete.LeastVisitedBonusExplorer.activate

LeastVisitedBonusExplorer.activate(qvalues=None, *args, **kwargs)

Turn on exploration mode.

If it is predicted that only states with rewards less than the threshold can be reached then the agent goes into exploration mode.

Parameters:

qvalues : dict

The qvalues for all actions from the current state