rickhenderson

Ran the random agent on FrozenLake-v0 with a total of 10000 episodes since there were not enough real results with only 100 episodes.

rickhenderson

Ran random_agent.py from the examples folder in the default gym installation.

rickhenderson

Using sample() on Berzerk-v0

This experiment was created by using the sample code provided in the docs at https://gym.openai.com/docs and I simply changed the environment to Berzerk-v0 and changed the number of episodes for 100 for proper scoring.

Reproduction

1. Follow the demo at https://gym.openai.com/docs
2. Use env = gym.make('Berzerk-v0') for the enviroment
3. Set the episodes to a minimum of 100.

-OR- Use the code from here:

rickhenderson

# mandelbrot_vectorized.R
# Myles Harrison
# http://www.everydayanalytics.ca

# parameters
cols <- colorRampPalette(c("blue","yellow","red","black"))(11)
xmin = -2
xmax = 2
nx = 500
ymin = -1.5

rickhenderson

# mineral_overlay.py
# Display minerals using Mineral overlay using Minecraft Overviewer
# Updated May 6, 2016

outputdir = "c:/maps/devil"
worlds["Devil Land"] = "%appdata%/.minecraft/saves/Devil Land"

# Rendermodes for lighting include: normal, night, smooth_lighting, smooth_night, nether_smooth_lighting
renders["dlday"] = {
    'world': "Devil Land",