TODO: DiscreteDP: Allow beta=1

[oyamad]

Re item 3 in #185 (comment):

• Let DiscreteDP accept beta=1
• Disable evaluate_policy, policy_iteration, and modified_policy_iteration EDIT: and value_iteration: let them raise a NotImplementedError if beta=1
• Update the docstring

Also write a notebook on Miranda-Fackler Section 7.6.6.
EDIT: Done by QuantEcon/QuantEcon.notebooks#37.