environment gif

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

Name Value
Actions Discrete
Agents 121
Parallel API Yes
Manual Control No
Action Shape (5),(9)
Action Values Discrete(5),(9)
Observation Shape (3,3,5), (9,9,5)
Observation Values [0,2]
Import pettingzoo.magent import tiger_deer_v3
Agents agents= [ deer_[0-100], tiger_[0-19] ]

Agent Environment Cycle

environment aec diagram


In tiger-deer, there are a number of tigers who are only rewarded for teaming up to take down the deer (two tigers must attack a deer in the same step to receive reward). If they do not eat the deer, they will slowly lose 0.1 HP each turn until they die. If they do eat the deer they regain 8 health (they have 10 health to start). At the same time, the deer are trying to avoid getting attacked. Deer start with 5 HP, lose 1 HP when attacked, and regain 0.1 HP each turn. Deer should run from tigers and tigers should form small teams to take down deer.

Action Space

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

Tiger action space: [do_nothing, move_4, attack_4]

Deer action space: [do_nothing, move_4]


Tiger’s reward scheme is:

Deer’s reward scheme is:

Observation space

The observation space is a 3x3 map with 5 channels for deer and 9x9 map with 5 channels for tigers, which are (in order):

name number of channels
obstacle/off the map 1
my_team_presence 1
my_team_hp 1
other_team_presence 1
other_team_hp 1
binary_agent_id(extra_features=True) 10
one_hot_action(extra_features=True) 5 Deer/9 Tiger
last_reward(extra_features=True) 1


tiger_deer_v3.env(map_size=45, minimap_mode=False, tiger_step_recover=-0.1, deer_attacked=-0.1, max_cycles=500, extra_features=False)

map_size: Sets dimensions of the (square) map. Increasing the size increases the number of agents. Minimum size is 10.

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).

tiger_step_recover: Amount of health a tiger gains/loses per turn (tigers have health 10 and get health 8 from killing a deer)

deer_attacked: Reward a deer gets for being attacked

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

extra_features: Adds additional features to observation (see table). Default False