CoreRandom

---

Abstract base class for all implemented random processes.

Do not use this abstract base class directly but instead use one of the concrete random processes implemented.

To implement your own random processes, you have to implement the following methods:

• decay
• sample
• reset

Methods:

.reset

.reset()

---

Reset random state.

.decay

.decay()

---

decay

.sample

.sample()

---

Sample random state.

Returns

sample (abstract): Random state.

---

GaussianRandom

GaussianRandom(
   mu = 0.0, size = 2, sigma = 0.1, sigma_min = 0.01, decay_steps = 200000
)

---

Gaussian Noise

Arguments

mu (float): . size (int): . sigma (float): . sigma_min (float): . decay_steps (int): .

Methods:

.reset

.reset()

---

Remember the transaction.

Accepts a state, action, reward, next_state, terminal transaction.

Arguments

transaction (abstract): state, action, reward, next_state, terminal transaction.

.decay

.decay()

---

Remember the transaction.

Accepts a state, action, reward, next_state, terminal transaction.

Arguments

transaction (abstract): state, action, reward, next_state, terminal transaction.

.sample

.sample()

---

Remember the transaction.

Accepts a state, action, reward, next_state, terminal transaction.

Arguments

transaction (abstract): state, action, reward, next_state, terminal transaction.

---

OrnsteinUhlenbeck

OrnsteinUhlenbeck(
   dt = 1.0, mu = 0.0, size = 2, sigma = 0.1, theta = 0.15, sigma_min = 0.01,
   decay_steps = 200000
)

---

Ornstein Uhlenbeck Process

Arguments

dt (float): . mu (float): . size (int): . sigma (float): . theta (float): . sigma_min (float): . decay_steps (int): .

Methods:

.reset

.reset()

---

Remember the transaction.

Accepts a state, action, reward, next_state, terminal transaction.

Arguments

transaction (abstract): state, action, reward, next_state, terminal transaction.

.decay

.decay()

---

Remember the transaction.

Accepts a state, action, reward, next_state, terminal transaction.

Arguments

transaction (abstract): state, action, reward, next_state, terminal transaction.

.sample

.sample()

---

Remember the transaction.

Accepts a state, action, reward, next_state, terminal transaction.

Arguments

transaction (abstract): state, action, reward, next_state, terminal transaction.