CODEWITHSUNDEEP
development-environmentreinforcement-learningopenai-gym
Savinay_
Savinay_

Reputation: 41

Custom environments in OpenAI-Gym

I want to try out reinforcement learning in Python. But I want to create a custom environment with my own States and Rewards.

Possibly I want to use Time-Series Data in this(nothing like Stocks though). I want my system to learn behaviors according to time, given an environment.

So I would first start by making a custom environment, Can someone help me in that ?

Upvotes: 2

Views: 4998

Answers (1)

Vishma Dias
Vishma Dias

Reputation: 700

The OpenAI gym environment registration process can be found in the gym docs here.

You could also check out this example custom environment and this stackoverflow issue for further information.

Possibly I want to use Time-Series Data in this(nothing like Stocks though). I want my system to learn behaviors according to time, given an environment.

If you want the agent to learn w.r.t. time then include time-step as a variable in your state. Then your agent will be able to distinguish between two distinct states differ from time-step.

Upvotes: 2

Related Questions