Skip to main content

Environments for the Statistical Reinforcement Learning project

Project description

Statistical Reinforcement Learning Environments

Environments for Statistical Reinforcement Learning

Installation

pip install statisticalRL-environments

Test

python fulldemo.py

List of environments:

MABS:
    mab-bernoulli
MDPs:
    random-rich
    ergodic-random-rich
    random-12
    random-small
    random-small-sparse
    random-100
    three-state
    nasty
    river-swim-6
    ergo-river-swim-6
    ergo-river-swim-25
    river-swim-25
GRIDWORLD MDPs:
    grid-random-1616
    grid-random-1212
    grid-random-88
    grid-2-room
    grid-4-room

The library contains several stochastic MDPs, that is for which transition and reward functions are not deterministic.

Randomly generated environments:

Furthermore, the library supports random generations of MDPs. The environments containing "random" in their name are randomly generated MDPs, that is with randomly generated transition and reward (stochastic) functions. For reproducibility reasons, they are generated with a given seed.

You can modify them in the registration list "registerStatisticalRLenvironments" available in the init file of lib. In this list, you will find environments registered with specific parameters, including the (last) parameter "seed" which is used to generate the random transition and reward functions.

"random-small" : lambda x: registerRandomMDP(nbStates=3, nbActions=4, maxProportionSupportTransition=0.5, maxProportionSupportReward=0.4, maxProportionSupportStart=0.1, minNonZeroProbability=0.15, minNonZeroReward=0.25, rewardStd=0.1,seed=5),

Note: fixing this seed (here sedd=5) does not prevent the transitions/rewards to be stochastic, as they are using another random number generating process.

Rendering

Each type of environment comes with different renderers, including the null renderer that displays nothing.

Text rendering:

This rendering is available for all environment types.

MAB Text rendering:

alt text

MDP Text rendering:

alt text

Gridworld-MDP Text rendering

alt text

Graph rendering:

This rendering is available for MDPs. On top of the visual display, it captures a screenshot png at each time step. This may be slow for large MDPs.

alt text

alt text

alt text

Grid-world rendering:

This rendering is available for Gridworld-MDPs. On top of the visual display, it captures a screenshot png at each time step.

alt text

alt text

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

statisticalrl_environments-2.2507.tar.gz (21.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

statisticalrl_environments-2.2507-py3-none-any.whl (29.4 kB view details)

Uploaded Python 3

File details

Details for the file statisticalrl_environments-2.2507.tar.gz.

File metadata

File hashes

Hashes for statisticalrl_environments-2.2507.tar.gz
Algorithm Hash digest
SHA256 f7ffcd98f12530d626c3c8540528281312d484902f2473c73d58b0689c924386
MD5 fcd2e56335a8c0c2ffc47da0c5ad2ec2
BLAKE2b-256 4e4629c65ebe2a35053b4d434f892084ba5c5116b2d7947a0256800f190df08d

See more details on using hashes here.

File details

Details for the file statisticalrl_environments-2.2507-py3-none-any.whl.

File metadata

File hashes

Hashes for statisticalrl_environments-2.2507-py3-none-any.whl
Algorithm Hash digest
SHA256 6bd941ea312ee3fa94509c136e4cf733828c2808e527bfdf6cf28bf14a8caafe
MD5 141f839f5b295cfef53950ad37ab543a
BLAKE2b-256 db631efd7e351916d00767f3077f4b4ada6b52adce68318ee67b18da4c58258f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page