markusdumke/reinforcelearn

Issues

• Some help setting up a toy example: where to define the action space?

  #37 opened 4 years ago by b-rodrigues
  0

• Errors in the Agents vignette

  #32 opened 6 years ago by russellcameronthomas
  7

• an error reported using kera neural network model

  #36 opened 5 years ago by zh-zhang1984
  0

• interact with OpenAI gyms

  #35 opened 5 years ago by OGuggenbuehl
  0

• Fix issues with new R Version 3.6.0

  #33 opened 5 years ago by markusdumke
  1

• keras training is too slow

  #28 opened 7 years ago by berndbischl
  2

• qlearning et all: no arg check on class for first arg. that must be an RL environment

  #27 opened 7 years ago by berndbischl
  0

• fun.approx = "neural.network" uninformative error

  #25 opened 7 years ago by berndbischl
  0

• ?qlearning et al are missing precise docs for return value

  #19 opened 7 years ago by berndbischl
  0

• env printer seems to be not informative enough

  #20 opened 7 years ago by berndbischl
  1

• can we visualize policy steps and what happens during learning?

  #14 opened 7 years ago by berndbischl
  1

• it would be nice if states and actions have string names

  #16 opened 7 years ago by berndbischl
  1

• sarsa on mountaincar leads to uninformative error - because we have no func approx

  #17 opened 7 years ago by berndbischl
  0

• more uninformative error

  #26 opened 7 years ago by berndbischl
  0

• is there an R6 object to represent a policy?

  #15 opened 7 years ago by berndbischl
  1

• can we have a better startup procedure for ai gym?

  #18 opened 7 years ago by berndbischl
  1

• what happens in eg mountaincar if the car is not pushed to the top?

  #21 opened 7 years ago by berndbischl
  1

• how can i run my learned policy

  #22 opened 7 years ago by berndbischl
  1

• Error when calling `makeEnvironment` after start of Rstudio

  #13 opened 7 years ago by markusdumke
  5

• README: we should copy the install instructs for openai gym and keras to our readme

  #24 opened 7 years ago by berndbischl
  0

• qlearning et al: preprocessState docs

  #23 opened 7 years ago by berndbischl
  0

• Double Q-Learning not working

  #8 opened 7 years ago by markusdumke
  0

• Implement Policy Iteration and Value Iteration

  #5 opened 7 years ago by markusdumke
  0

• Expected Sarsa not giving the expected results on gridworld

  #6 opened 7 years ago by markusdumke
  0

• Implement MC control

  #7 opened 7 years ago by markusdumke
  0

• Implement algorithms with value-function approximation

  #10 opened 7 years ago by markusdumke
  0

• MonteCarlo Control with Exploring Starts not working

  #11 opened 7 years ago by markusdumke
  0

• Test for Sarsa fails

  #12 opened 7 years ago by markusdumke
  0

• Implement lambda version of sarsa

  #9 opened 7 years ago by markusdumke
  0

• TD not working

  #4 opened 7 years ago by markusdumke
  0

• predictMC not working

  #3 opened 7 years ago by markusdumke
  0

• implement gridworld as R6 class

  #1 opened 7 years ago by markusdumke
  0

• fix bug in gridworld.R

  #2 opened 7 years ago by markusdumke
  0