Improper Assertion Statement.

[ryanpeach]

assert self.max_iter >= 10000, "'n_iter' should be greater than 10000."

This line in QLearning is inconvenient and arbitrary. It should be either removed or turned into a warning.

I am interested in hacking these classes, for instance, such that they can iteratively learn as the Reward matrix changes. As such, small values of this value are necessary.