First order

Action space	Box((1,)), all actions in [-1, 1]
Observation space	Box((2,))
Tracked variable(s)	x
Episode length	100 steps (5 s at 0.05 s per step)
Import	from target_gym import FirstOrderSystem, FirstOrderParams
Cite as	first_order-v1

Action space

Actions are normalised to [-1, 1] and mapped onto the plant's real actuator range inside the environment.

#	meaning	min	max
0	input	-1	1

Observation space

2 values. Indices (0,) carry the tracked variable(s) that the reward scores.

Rewards

See the environment's compute_reward.

Every environment in this suite scores on one contract: the reward is (tracking terms, multiplied) x (1 - weighted costs), bounded in [0, 1], and reaches 1 only while the target is held exactly. See Reward shaping.

Starting state

reset samples the initial condition and the target; state has 4 fields.

Episode end

Termination. See check_is_terminal.

Truncation. After 100 steps.

Baselines

A tuned PID ships with this environment.

Arguments

parameter	default
delta_t	0.05
max_steps_in_episode	100
K	1
tau	0.5
u_min	-2
u_max	2
x_min	-3
precision_floor	0.006
x_max	3