Quadrapong

../../../_images/atari_quadrapong.gif

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

Import

from pettingzoo.atari import quadrapong_v4

Actions

Discrete

Parallel API

Yes

Manual Control

No

Agents

agents= ['first_0', 'second_0', 'third_0', 'fourth_0']

Agents

4

Action Shape

(1,)

Action Values

[0,5]

Observation Shape

(210, 160, 3)

Observation Values

(0,255)

Four player team battle.

Each player controls a paddle and defends a scoring area. However, this is a team game, and so two of the 4 scoring areas belong to the same team. So a given team must try to coordinate to get the ball away from their scoring areas towards their opponent’s. Specifically first_0 and third_0 are on one team and second_0 and fourth_0 are on the other.

Scoring a point gives your team +1 reward and your opponent team -1 reward.

Serves are timed: If the player does not serve within 2 seconds of receiving the ball, their team receives -1 points, and the timer resets. This prevents one player from indefinitely stalling the game, but also means it is no longer a purely zero sum game.

Official Video Olympics manual

Environment parameters

Environment parameters are common to all Atari environments and are described in the base Atari documentation .

Action Space (Minimal)

In any given turn, an agent can choose from one of 6 actions.

Action

Behavior

0

No operation

1

Fire

2

Move up

3

Move right

4

Move left

5

Move down

Version History

  • v4: Minimal Action Space (1.18.0)

  • v3: No action timer (1.9.0)

  • v1: Breaking changes to entire API (1.4.0)

  • v2: Fixed quadrapong rewards (1.2.0)

  • v0: Initial versions release (1.0.0)

Usage

AEC

from pettingzoo.atari import quadrapong_v4

env = quadrapong_v4.env(render_mode="human")
env.reset(seed=42)

for agent in env.agent_iter():
    observation, reward, termination, truncation, info = env.last()

    if termination or truncation:
        action = None
    else:
        # this is where you would insert your policy
        action = env.action_space(agent).sample()

    env.step(action)
env.close()

Parallel

from pettingzoo.atari import quadrapong_v4

env = quadrapong_v4.parallel_env(render_mode="human")
observations, infos = env.reset()

while env.agents:
    # this is where you would insert your policy
    actions = {agent: env.action_space(agent).sample() for agent in env.agents}

    observations, rewards, terminations, truncations, infos = env.step(actions)
env.close()

API

class pettingzoo.atari.quadrapong.quadrapong.raw_env(**kwargs)[source]