mirror of
https://github.com/illiumst/marl-factory-grid.git
synced 2025-05-23 07:16:44 +02:00
Refactoring of movement logic and parallel collision checks
This commit is contained in:
parent
ebd8f46a16
commit
c539e5dddd
@ -31,19 +31,27 @@ class BaseFactory:
|
|||||||
# Returns State, Reward, Done, Info
|
# Returns State, Reward, Done, Info
|
||||||
return self.state, 0, self.done, {}
|
return self.state, 0, self.done, {}
|
||||||
|
|
||||||
|
def additional_actions(self, agent_i, action) -> ((int, int), bool):
|
||||||
|
raise NotImplementedError
|
||||||
|
|
||||||
def step(self, actions):
|
def step(self, actions):
|
||||||
actions = [actions] if isinstance(actions, int) else actions
|
actions = [actions] if isinstance(actions, int) else actions
|
||||||
assert isinstance(actions, list), f'"actions has to be in [{int, list}]'
|
assert isinstance(actions, list), f'"actions has to be in [{int, list}]'
|
||||||
self.steps += 1
|
self.steps += 1
|
||||||
r = 0
|
r = 0
|
||||||
|
|
||||||
collision_vecs = np.zeros((self.n_agents, self.state.shape[0])) # n_agents x n_slices
|
|
||||||
actions = list(enumerate(actions))
|
actions = list(enumerate(actions))
|
||||||
random.shuffle(actions)
|
random.shuffle(actions)
|
||||||
|
|
||||||
for agent_i, action in actions:
|
for agent_i, action in actions:
|
||||||
new_pos, collision_vec, did_collide = self.move_or_colide(agent_i, action)
|
if action <= 8:
|
||||||
collision_vecs[agent_i] = collision_vec
|
pos, did_collide = self.move_or_colide(agent_i, action)
|
||||||
|
else:
|
||||||
|
pos, did_collide = self.additional_actions(agent_i, action)
|
||||||
|
actions[agent_i] = (pos, did_collide)
|
||||||
|
|
||||||
|
collision_vecs = np.zeros((self.n_agents, self.state.shape[0])) # n_agents x n_slices
|
||||||
|
for agent_i, action in enumerate(actions):
|
||||||
|
collision_vecs[agent_i] = self.check_collisions(agent_i, *action)
|
||||||
|
|
||||||
reward, info = self.step_core(collision_vecs, actions, r)
|
reward, info = self.step_core(collision_vecs, actions, r)
|
||||||
r += reward
|
r += reward
|
||||||
@ -51,22 +59,32 @@ class BaseFactory:
|
|||||||
self.done = True
|
self.done = True
|
||||||
return self.state, r, self.done, info
|
return self.state, r, self.done, info
|
||||||
|
|
||||||
|
def check_collisions(self, agent_i, pos, valid):
|
||||||
|
pos_x, pos_y = pos
|
||||||
|
collisions_vec = self.state[:, pos_x, pos_y].copy() # "vertical fiber" at position of agent i
|
||||||
|
collisions_vec[h.AGENT_START_IDX + agent_i] = h.IS_FREE_CELL # no self-collisions
|
||||||
|
if valid:
|
||||||
|
pass
|
||||||
|
else:
|
||||||
|
collisions_vec[h.LEVEL_IDX] = h.IS_OCCUPIED_CELL
|
||||||
|
return collisions_vec
|
||||||
|
|
||||||
def move(self, agent_i, old_pos, new_pos):
|
def move(self, agent_i, old_pos, new_pos):
|
||||||
(x, y), (x_new, y_new) = old_pos, new_pos
|
(x, y), (x_new, y_new) = old_pos, new_pos
|
||||||
self.state[agent_i + h.AGENT_START_IDX, x, y] = h.IS_FREE_CELL
|
self.state[agent_i + h.AGENT_START_IDX, x, y] = h.IS_FREE_CELL
|
||||||
self.state[agent_i + h.AGENT_START_IDX, x_new, y_new] = h.IS_OCCUPIED_CELL
|
self.state[agent_i + h.AGENT_START_IDX, x_new, y_new] = h.IS_OCCUPIED_CELL
|
||||||
|
|
||||||
def move_or_colide(self, agent_i, action) -> ((int, int), bool):
|
def move_or_colide(self, agent_i, action) -> ((int, int), bool):
|
||||||
old_pos, new_pos, collision_vec, did_collide = h.check_agent_move(state=self.state,
|
old_pos, new_pos, valid = h.check_agent_move(state=self.state,
|
||||||
dim=agent_i + h.AGENT_START_IDX,
|
dim=agent_i + h.AGENT_START_IDX,
|
||||||
action=action)
|
action=action)
|
||||||
if not did_collide:
|
if valid:
|
||||||
# Does not collide width level boundrys
|
# Does not collide width level boundrys
|
||||||
self.move(agent_i, old_pos, new_pos)
|
self.move(agent_i, old_pos, new_pos)
|
||||||
return new_pos, collision_vec, did_collide
|
return new_pos, valid
|
||||||
else:
|
else:
|
||||||
# Agent seems to be trying to collide in this step
|
# Agent seems to be trying to collide in this step
|
||||||
return old_pos, collision_vec, did_collide
|
return old_pos, valid
|
||||||
|
|
||||||
@property
|
@property
|
||||||
def free_cells(self) -> np.ndarray:
|
def free_cells(self) -> np.ndarray:
|
||||||
|
@ -30,6 +30,6 @@ class SimpleFactory(BaseFactory):
|
|||||||
if __name__ == '__main__':
|
if __name__ == '__main__':
|
||||||
import random
|
import random
|
||||||
factory = SimpleFactory(n_agents=1, max_dirt=8)
|
factory = SimpleFactory(n_agents=1, max_dirt=8)
|
||||||
random_actions = [random.randint(0, 8) for _ in range(200)]
|
random_actions = [random.randint(0, 7) for _ in range(200)]
|
||||||
for action in random_actions:
|
for action in random_actions:
|
||||||
state, r, done, _ = factory.step(action)
|
state, r, done, _ = factory.step(action)
|
||||||
|
@ -64,16 +64,7 @@ def check_agent_move(state, dim, action):
|
|||||||
or y_new >= agent_slice.shape[0]
|
or y_new >= agent_slice.shape[0]
|
||||||
)
|
)
|
||||||
|
|
||||||
if valid:
|
return (x, y), (x_new, y_new), valid
|
||||||
collisions_vec = state[:, x_new, y_new].copy() # "vertical fiber" at position of agent i
|
|
||||||
collisions_vec[dim] = IS_FREE_CELL # no self-collisions
|
|
||||||
pass
|
|
||||||
else:
|
|
||||||
collisions_vec = state[:, x, y].copy() # "vertical fiber" at position of agent i
|
|
||||||
collisions_vec[dim] = IS_FREE_CELL # no self-collisions
|
|
||||||
collisions_vec[LEVEL_IDX] = IS_OCCUPIED_CELL
|
|
||||||
did_collide = collisions_vec.sum(0) != IS_FREE_CELL
|
|
||||||
return (x, y), (x_new, y_new), collisions_vec, did_collide
|
|
||||||
|
|
||||||
|
|
||||||
if __name__ == '__main__':
|
if __name__ == '__main__':
|
||||||
|
Loading…
x
Reference in New Issue
Block a user