Skip to content

Tresta contains Heuristics, Reinforcement Learning, Graph based Learning related Implementation

Notifications You must be signed in to change notification settings

HSaurabh0919/tresta

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

39 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Building beautiful RL products, related to supply chain, gaming, Economics and Finance Domain!

  1. Cartpole Solved using Stable-baselines Implemenation.
  2. Cartpole using A2C: Synchronous Advantage Actor Critic Implementation.
  3. DDQN implementation of Cart-Pole can be found here and also has comparison related to various epsilon-decay methods.

😊😊😊😊

Plots

  1. Comparison of effect of epsilon-decay approach on DDQN implementation for Cartpole environment.
  • 1.1. Linear Epsilon Decay

Linear Epsilon Decay

  • 1.2. Exponential Epsilon Decay

Exponential Epsilon Decay

  • 1.3. Exponential-Squared Epsilon Decay

Exponential Squared Epsilon Decay

  • 1.4. All epslion values

 Decay

Customized Environments

  1. 2D Maze Game

image

REFERENCES:

  1. Cartpole custom A2C Implementation has been adapted from here.
  2. DDQN custom Implementation has been majorly inspired from here.
  3. RL trading has been adapted from here.
  4. Trading environment has been used from here.

About

Tresta contains Heuristics, Reinforcement Learning, Graph based Learning related Implementation

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published