12. Exercise 4: Manual Training Steps

This problem is designed to help you understand the reinforcement learning process from a financial perspective.

Scenario Definition:  Above we have depicted a reinforcement learning environment. Imagine there is a trading bot who is trying to make a profit by trading this stock. We want to use three sources of information in this environment - the stock close price, the 20 day moving average of the stock close price, and the volume of stocks being traded. Values for these variables are provided for specific time steps below.

Objective: For time steps 0 to 6, define the state, action, next state, and reward, utilizing the definitions from Exercise 2.

• HINT: the format of the state representation vector should be [close price, moving average, volume].