Ant

ant

This Environment is part of MaMuJoCo environments. Please read that page first for general information. The task is Gymansium’s MuJoCo/Ant.

Defaults

env = mamujoco_v1.parallel_env("Ant", None)

Agents

agents= ['agent_0']

Number of Agents

1

Action Spaces

{'agent_0' : Box(-1, 1, (8,), float32)}

Part partition

[(hip4, ankle4, hip1, ankle1, hip2, ankle2, hip3, ankle3)]

Action Space

The shape of the action space depends on the partitioning. The partitioning has the following form:

if partitioning is None:

../../../_images/ant.png

Instantiate

env = mamujoco_v1.parallel_env("Ant", None)

Agents

agents= ['agent_0']

Number of Agents

1

Action Spaces

{'agent_0' : Box(-1, 1, (8,), float32)}

Part partition

[(hip4, ankle4, hip1, ankle1, hip2, ankle2, hip3, ankle3)]

If partitioning, is None then the environment contains a single agent with the same action space as Gymansium’s MuJoCo/Ant.

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and back right hip

-1

1

hip_4 (right_back_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the back right two links

-1

1

angle_4 (right_back_leg)

hinge

torque (N m)

2

Torque applied on the rotor between the torso and front left hip

-1

1

hip_1 (front_left_leg)

hinge

torque (N m)

3

Torque applied on the rotor between the front left two links

-1

1

angle_1 (front_left_leg)

hinge

torque (N m)

4

Torque applied on the rotor between the torso and front right hip

-1

1

hip_2 (front_right_leg)

hinge

torque (N m)

5

Torque applied on the rotor between the front right two links

-1

1

angle_2 (front_right_leg)

hinge

torque (N m)

6

Torque applied on the rotor between the torso and back left hip

-1

1

hip_3 (back_leg)

hinge

torque (N m)

7

Torque applied on the rotor between the back left two links

-1

1

angle_3 (back_leg)

hinge

torque (N m)

if partitioning == “2x4”: # neighboring legs together (front and back)

../../../_images/ant_2x4.png

Instantiate

env = mamujoco_v1.parallel_env("Ant", "2x4")

Agents

agents= ['agent_0', 'agent_1']

Number of Agents

2

Action Spaces

{'agent_0' : Box(-1, 1, (4,), float32), 'agent_1' : Box(-1, 1, (4,), float32)}

Part partition

[(hip1, ankle1, hip2, ankle2), (hip3, ankle3, hip4, ankle4)]

The environment is partitioned in 2 parts, the front part (containing the front legs) and the back part (containing the back legs).

Agent 0 action space (front legs)

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and front left hip

-1

1

hip_1 (front_left_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the front left two links

-1

1

angle_1 (front_left_leg)

hinge

torque (N m)

2

Torque applied on the rotor between the torso and front right hip

-1

1

hip_2 (front_right_leg)

hinge

torque (N m)

3

Torque applied on the rotor between the front right two links

-1

1

angle_2 (front_right_leg)

hinge

torque (N m)

Agent 1 action space (back legs)

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and back right hip

-1

1

hip_4 (right_back_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the back right two links

-1

1

angle_4 (right_back_leg)

hinge

torque (N m)

2

Torque applied on the rotor between the torso and back left hip

-1

1

hip_3 (back_leg)

hinge

torque (N m)

3

Torque applied on the rotor between the back left two links

-1

1

angle_3 (back_leg)

hinge

torque (N m)

if partitioning == “2x4d”: # diagonal legs together

../../../_images/ant_2x4d.png

Instantiate

env = mamujoco_v1.parallel_env("Ant", "2x4d")

Agents

agents= ['agent_0', 'agent_1']

Number of Agents

2

Action Spaces

{'agent_0' : Box(-1, 1, (4,), float32), 'agent_1' : Box(-1, 1, (4,), float32)}

Part partition

[(hip1, ankle1, hip4, ankle4), (hip2, ankle2, hip3, ankle3)]

The environment is partitioned in 2 parts, split diagonally.

Agent 0 action space

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and front left hip

-1

1

hip_1 (front_left_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the front left two links

-1

1

angle_1 (front_left_leg)

hinge

torque (N m)

2

Torque applied on the rotor between the torso and back right hip

-1

1

hip_4 (right_back_leg)

hinge

torque (N m)

3

Torque applied on the rotor between the back right two links

-1

1

angle_4 (right_back_leg)

hinge

torque (N m)

Agent 1 action space

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and front right hip

-1

1

hip_2 (front_right_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the front right two links

-1

1

angle_2 (front_right_leg)

hinge

torque (N m)

2

Torque applied on the rotor between the torso and back left hip

-1

1

hip_3 (back_leg)

hinge

torque (N m)

3

Torque applied on the rotor between the back left two links

-1

1

angle_3 (back_leg)

hinge

torque (N m)

if partitioning == “4x2”:

../../../_images/ant_4x2.png

Instantiate

env = mamujoco_v1.parallel_env("Ant", "4x2")

Agents

agents= ['agent_0', 'agent_1', 'agent_2', 'agent_3']

Number of Agents

4

Action Spaces

{'agent_0' : Box(-1, 1, (2,), float32), 'agent_1' : Box(-1, 1, (2,), float32)}, 'agent_2' : Box(-1, 1, (2,), float32), 'agent_3' : Box(-1, 1, (2,), float32)},

Part partition

[(hip1, ankle1), (hip2, ankle2), (hip3, ankle3), (hip4, ankle4)]

The environment is partitioned in 4 parts, with each part corresponding to a leg of the ant.

Agent 0 action space (front left leg)

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and front left hip

-1

1

hip_1 (front_left_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the front left two links

-1

1

angle_1 (front_left_leg)

hinge

torque (N m)

Agent 1 action space (front right leg)

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and front right hip

-1

1

hip_2 (front_right_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the front right two links

-1

1

angle_2 (front_right_leg)

hinge

torque (N m)

Agent 2 action space (right left leg)

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

2

Torque applied on the rotor between the torso and back left hip

-1

1

hip_3 (back_leg)

hinge

torque (N m)

3

Torque applied on the rotor between the back left two links

-1

1

angle_3 (back_leg)

hinge

torque (N m)

Agent 3 action space (right back leg)

Num

Action

Control Min

Control Max

Name (in corresponding XML file)

Joint

Unit

0

Torque applied on the rotor between the torso and back right hip

-1

1

hip_4 (right_back_leg)

hinge

torque (N m)

1

Torque applied on the rotor between the back right two links

-1

1

angle_4 (right_back_leg)

hinge

torque (N m)

Observation Space

Observation Categories

Default local_categories

[["qpos", "qvel"], ["qpos"]]

Default global_categories

("qpos", "qvel")

Supported observation categories

"qpos", "qvel", "cfrc_ext"

Besides the local observation of each agent (which depend on their parts of the agent, the observation categories and the observation depth), each agent also observes the position and velocity items of the ant’s torso. See more at the Gymnasium’s Ant.

Rewards

All agents receive the same Gymnasium’s Ant reward.

Starting state

The starting state of the environment is the same as Gymnasium’s Ant.

Episode End

All agent terminate and truncate at the same time given the same conditions as Gymnasium’s Ant.

Version History

  • v1:

    • Now based on Gymnasium/MuJoCo-v5 instead of Gymnasium/MuJoCo-v4 (https://github.com/Farama-Foundation/Gymnasium/pull/572).

    • Now observes local_categories of cfrc_ext by default (same as Gymnasium/MuJoCo-v5/Ant).

    • Renamed global node torso → root.

  • v0: Initial version release, uses Gymnasium.MuJoCo-v4, and is a fork of the original MaMuJoCo schroederdewitt/multiagent_mujoco. Changes from the original MaMuJoCo (schroederdewitt/multiagent_mujoco):

    • Fixed diagonal factorization (“2x4d”) not being diagonal.

    • Fixed Global observations (The Ant’s Torso: rootx, rooty, rootz) not being observed.

    • Changed action ordering to be same as Gymnasium/MuJoCo/Ant