Simple Q Learning/SARSA Homework Assignment

[zsunberg]

I would like to have students do a basic homework assignment with CommonRLInterface. This will have two parts:

1. I will give them a grid world environment with a clear visualization so that they can see what is going on and make sure their algorithm works.
2. I will give them a mystery discrete environment (with obfuscated code) without a visualization so that they can experience it like the agent does with meaningless observations and only a reward signal.

For part 1, the optional interface parts I need are:

1. observations(env) observations(env) #10
2. A way to render the environment.

For part 2, just the required interface will suffice.