02. Exercise 1: Animal Habitat

This problem is designed to help you understand the basic components of reinforcement learning from a real-world perspective.

**Scenario Definition: **Below we have depicted a reinforcement learning environment. There is one human who is trying to survive in the environment. The goal of the human in this environment is to find food and avoid predators. There are three food sources in the environment- the deer, the fish, and the bird. There is one predator in the environment - the bear.

**Relevant Assumptions: **

Assume that the human can move one step at a time.

Assume that the food and predator cannot move.

Assume the tree and lake are not obstacles - the human can “walk through” them to the food sources.

**Objective: **Your task is to identify the main reinforcement learning components for the scenario defined above.

Agent:

1. Who is the agent in this scenario?

**State Representation: **

1. What information needs to be included in the state for the agent to make informed decisions?
2. Describe the components of the state.

Action Space:

1. List the possible actions the agent can take in the environment.

Reward:

1. Define the rewards that will guide the agent’s behavior.
2. Specify the rewards for finding food, being caught by a predator, and any other relevant events.