This is an experiment to use RL on stocks trading. It's only for fun, we won't be rich with this (or yes?).
We have a stock market and many agents, we can use dummy agents and actor-critic agents. You can implement your own agent, training it and put it to compete with the other agents.
In the market there's money and only one kind of product, the agents start with the same money and quantity of product.
The goal is earn money, the reward is the money that the agent has earned (negative if it lost money).
Thanks for the book that I can study and understand about Reinforcement Learning.
- Maxim Lapan. Deep Reiforcement Learning Hands-On
Music: https://www.bensound.com
- Make a league where the best agents go to the next competition.
- Make a recurrent neural network (use LSTM).