Rock Paper Scissors

This environment is part of the Classic environments. Please read that page first for general information.


Possible Agents	(‘0’, ‘1’)
Action Spaces	{‘0’: Discrete(3), ‘1’: Discrete(3)}
Observation Spaces	{‘0’: Discrete(3), ‘1’: Discrete(3)}
Symmetric	True
Import	`posggym.make("RockPaperScissors-v0")`

The Rock Paper Scissors Environment.

This is the classic game of rock, paper, scissors (RPS).

This scenario involves two agents. Each step both agents choose an action out of ‘ROCK’, ‘PAPER’ or ‘SCISSORS’ and are rewarded based on the actions taken in comparison to their opponent.

Possible Agents

The environment supports two agents: ‘0’ and ‘1’. Both agents are always active in the environment.

State Space

There is only a single state in RPS: state 0.

Action Space

Both agents have three available actions: ROCK=0, PAPER=1, SCISSORS=2.

Observation Space

Agents observe the last action played by their opponent: ROCK=0, PAPER=1, SCISSORS=2

Rewards

Agents are rewarded based on the following pay-off matrix (shows pay-off for row agent):

	ROCK	PAPER	SCISSORS
ROCK	0	-1	1
PAPER	1	0	-1
SCISSORS	-1	1	0

Dynamics

There is only a single state so the transition function is the identity function.

Starting State

There is only a single state, which the environment always starts and remains in.

Episode End

By default episodes continue infinitely long. To set a step limit, specify max_episode_steps when initializing the environment with posggym.make.

Arguments

No additional arguments are currently supported during construction.

Version History

v0: Initial version