environment gif

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

Name Value
Actions Discrete
Agents 162
Parallel API Yes
Manual Control No
Action Shape (21)
Action Values Discrete(21)
Observation Shape (13,13,41)
Observation Values [0,2]
Import pettingzoo.magent import battle_v2
Agents agents= [red_[0-80], blue_[0-80]]

Agent Environment Cycle

environment aec diagram


A large-scale team battle. Agents are rewarded for their individual performance, and not for the performance of their neighbors, so coordination is difficult. Agents slowly regain HP over time, so it is best to kill an opposing agent quickly. Specifically, agents have 10 HP, are damaged 2 HP by each attack, and recover 0.1 HP every turn.

Like all MAgent environments, agents can either move or attack each turn. An attack against another agent on their own team will not be registered.

Action space

Key: move_N means N separate actions, one to move to each of the N nearest squares on the grid.

Action options: [do_nothing, move_12, attack_8]


Reward is given as:

If multiple options apply, rewards are added.

Observation space

The observation space is a 13x13 map with 41 channels, which are (in order):

name number of channels
obstacle/off the map 1
my_team_presence 1
my_team_hp 1
my_team_minimap 1
other_team_presence 1
other_team_hp 1
other_team_minimap 1
binary_agent_id 10
one_hot_action 21
last_reward 1
agent_position 2


battle_v2.env(map_size=45, minimap_mode=True, step_reward=-0.005, dead_penalty=-0.1, attack_penalty=-0.1, attack_opponent_reward=0.2, max_cycles=1000)

map_size: Sets dimensions of the (square) map. Increasing the size increases the number of agents. Minimum size is 12.

minimap_mode: Turns on global minimap observations. These observations include your and your opponents piece densities binned over the 2d grid of the observation space. Also includes your agent_position, the absolute position on the map (rescaled from 0 to 1).

step_reward: reward after every step

dead_penalty: reward when killed

attack_penalty: reward when attacking anything

attack_opponent_reward: reward added for attacking an opponent

max_cycles: number of frames (a step for each agent) until game terminates