Professional Documents
Culture Documents
SOLUTIONS:
Fully observable: The agent needs to receive all the information from the
enviroment at any time in order to take action to run, jump or not to jump over the
bar.
Single agent: The jumping and running action of the agent is performed by itself,
there is no interaction between other agents.
Stochastic: Jump bar maybe dropped, the height at which the agent jumps can be
different, sometimes without passing the bar.
Sequential: The current decisions could affect all future decisions. The agent needs
to run first, then jump and finally stop. Of course, their order cannot be reversed.