PrioritizedBuffer¶
python_motion_planning.local_planner.dqn.PrioritizedBuffer
¶
__init__(max_size, alpha=0.6, beta=0.4)
¶
Priority replay buffer.
push(*experience)
¶
Injecting an experience into the replay buffer.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
experience
|
tuple
|
five-element tuple including state, action, reward, next_state and done flag |
()
|
sample(batch_size)
¶
Sampling a batch of data.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
batch_size
|
int
|
the size of sampling batch |
required |
updatePriority(idx, td_error)
¶
Updating priorities based on temporal-difference