class TestMultiGrid(unittest.TestCase):
'''
Testing a toroidal MultiGrid
'''
torus = True
def setUp(self):
'''
Create a test non-toroidal grid and populate it with Mock Agents
'''
width = 3
height = 5
self.grid = MultiGrid(width, height, self.torus)
self.agents = []
counter = 0
for x in range(width):
for y in range(height):
for i in range(TEST_MULTIGRID[x][y]):
counter += 1
# Create and place the mock agent
a = MockAgent(counter, None)
self.agents.append(a)
self.grid.place_agent(a, (x, y))
def test_agent_positions(self):
'''
Ensure that the agents are all placed properly on the MultiGrid.
'''
for agent in self.agents:
x, y = agent.pos
assert agent in self.grid[x][y]
def test_neighbors(self):
'''
Test the toroidal MultiGrid neighborhood methods.
'''
neighborhood = self.grid.get_neighborhood((1, 1), moore=True)
assert len(neighborhood) == 8
neighborhood = self.grid.get_neighborhood((1, 4), moore=True)
assert len(neighborhood) == 8
neighborhood = self.grid.get_neighborhood((0, 0), moore=False)
assert len(neighborhood) == 4
neighbors = self.grid.get_neighbors((1, 4), moore=False)
assert len(neighbors) == 0
neighbors = self.grid.get_neighbors((1, 4), moore=True)
assert len(neighbors) == 5
neighbors = self.grid.get_neighbors((1, 1),
moore=False,
include_center=True)
assert len(neighbors) == 7
neighbors = self.grid.get_neighbors((1, 3), moore=False, radius=2)
assert len(neighbors) == 11
class TestMultiGrid(unittest.TestCase):
'''
Testing a toroidal MultiGrid
'''
torus = True
def setUp(self):
'''
Create a test non-toroidal grid and populate it with Mock Agents
'''
width = 3
height = 5
self.grid = MultiGrid(width, height, self.torus)
self.agents = []
counter = 0
for x in range(width):
for y in range(height):
for i in range(TEST_MULTIGRID[x][y]):
counter += 1
# Create and place the mock agent
a = MockAgent(counter, None)
self.agents.append(a)
self.grid.place_agent(a, (x, y))
def test_agent_positions(self):
'''
Ensure that the agents are all placed properly on the MultiGrid.
'''
for agent in self.agents:
x, y = agent.pos
assert agent in self.grid[x][y]
def test_neighbors(self):
'''
Test the toroidal MultiGrid neighborhood methods.
'''
neighborhood = self.grid.get_neighborhood((1, 1), moore=True)
assert len(neighborhood) == 8
neighborhood = self.grid.get_neighborhood((1, 4), moore=True)
assert len(neighborhood) == 8
neighborhood = self.grid.get_neighborhood((0, 0), moore=False)
assert len(neighborhood) == 4
neighbors = self.grid.get_neighbors((1, 4), moore=False)
assert len(neighbors) == 0
neighbors = self.grid.get_neighbors((1, 4), moore=True)
assert len(neighbors) == 5
neighbors = self.grid.get_neighbors((1, 1), moore=False,
include_center=True)
assert len(neighbors) == 7
neighbors = self.grid.get_neighbors((1, 3), moore=False, radius=2)
assert len(neighbors) == 11
class bacModel(Model):
'''world model for Eden Growth Model Simulation'''
def __init__(self, beginRad, splitChance, x=-1, y=-1, mut_rate=MUTATION_RATE):
self.running = True
self.num_agents = 0
self.schedule = RandomActivation(self)
self.grid = MultiGrid(WIDTH, HEIGHT, IS_TOROIDAL) #True for toroidal
#self.datacollector = DataCollector(
# model_reporters = {"Identifier": function_name}, #note no parentheses, just function name
# agent_reporter = {"Identifier2": function_name2})
if x == -1:
x = self.grid.width // 2
if y == -1:
y = self.grid.height // 2
#create subsequent agents
positions = self.grid.get_neighborhood((x,y), moore=False, radius=beginRad, include_center=True)
for coord in positions:
roll = random.random()
a = bacAgent(self.num_agents, self, False, splitChance)
self.num_agents += 1
self.schedule.add(a)
self.grid.place_agent(a, coord)
def step(self):
#self.datacollector.collect(self)
self.schedule.step()
class bacServerModel(Model):
################
###Deprecated###
################
def __init__(self, width, height, beginRad):
self.running = True
self.num_agents = width * height
self.schedule = RandomActivation(self)
self.grid = MultiGrid(width, height, IS_TOROIDAL) #True for toroidal
#self.datacollector = DataCollector(
# model_reporters = {"Identifier": function_name}, #note no parentheses, just function name
# agent_reporter = {"Identifier2": function_name2})
for x in range(self.grid.width):
for y in range(self.grid.height):
a = bacServerAgent(self.num_agents, self)
self.num_agents += 1
self.schedule.add(a)
self.grid.place_agent(a, (x,y))
x = self.grid.width // 2
y = self.grid.height // 2
#create subsequent agents
positions = self.grid.get_neighborhood((x,y), moore=False, radius=beginRad, include_center=True)
for coord in positions:
ag = self.grid.get_cell_list_contents([coord])[0]
ag.activate()
def step(self):
#self.datacollector.collect(self)
self.schedule.step()
def move(self):
'''
The random move function for the fish schools.
'''
neighbourhood = MultiGrid.get_neighborhood(self.model.grid, self.pos, True, False, 1)
new_pos = random.choice(neighbourhood)
MultiGrid.move_agent(self.model.grid, self, new_pos)
def fisherman_move(self):
'''
The random move function for the fisherman.
'''
neighbourhood = MultiGrid.get_neighborhood(self.model.grid, self.pos, True, False, 1)
new_pos = random.choice(neighbourhood)
while (new_pos[0] < self.model.no_fish_size) and (new_pos[1] < self.model.no_fish_size):
new_pos = random.choice(neighbourhood)
MultiGrid.move_agent(self.model.grid, self, new_pos)
class HoominWorld(Model):
LEFT = 0
RIGHT = 1
STRAIGHT = 2
FRIENDNODELOGNAME = "friendnodes"
TOTALMESSAGELOGNAME = "totalmessages"
STEPSTOCOMPLETIONLOGNAME = "stepstocompletion"
verbose = False
description = "A model of foot traffic and radio communication in an urban environment"
def __init__(self,
height=50,
width=50,
initial_hoomins=10,
logtag="default"):
super().__init__()
print("initializing ", settings.width, settings.height)
#map height and width
self.height = settings.height
self.width = settings.width
#logging framework
self.logger = Logging("logs", logtag)
self.logtag = logtag
self.G = nx.Graph()
#graph visualization
if not settings.runheadless:
plt.ion()
plt.show()
#ignore this. it does nothing
self.hoomin_level = 0
#road generation tuning
self.straightweight = settings.straightweight
self.leftweight = settings.leftweight
self.rightweight = settings.rightweight
self.initial_roads = settings.initial_roads
self.initial_road_seeds = settings.initial_road_seeds
self.gridspacing = settings.gridspacing
self.roadcurrentcoord = np.array((0, 0))
self.roaddir = np.array((1, 0))
self.roadset = set()
#home tuning options
self.homes_per_hoomins = 1
self.initial_homes = self.homes_per_hoomins * initial_hoomins
self.homeset = set()
#scatterbrain metrics
self.total_scattermessages = 0
self.global_scattermessages = 0
self.hoominzero_nodecount = 0
#hoomin tuning values
self.initial_hoomins = settings.initial_hoomins
self.schedule = RandomHoominActivation(self)
self.grid = MultiGrid(self.height, self.width, torus=True)
self.datacollector = DataCollector({
"Messages Exchanged":
lambda m: m.total_scattermessages,
"FriendGraph Node Count":
lambda m: m.hoominzero_nodecount
})
#initialize roads
for i in range(self.initial_road_seeds):
x = self.random.randrange(self.width)
y = self.random.randrange(self.height)
self.singleroad((x, y))
homelist = []
#initialize homes
for i in range(self.initial_homes):
road = self.random.sample(self.roadset, 1)
if len(road) > 0 and road[0] is not None:
neighbors = self.grid.get_neighborhood(road[0].pos, False,
True)
for neighbor in neighbors:
n = []
if len(self.grid.get_cell_list_contents(neighbor)) is 0:
n.append(neighbor)
if len(n) > 0:
homeblock = self.random.sample(n, 1)
home = Home(self.next_id(), homeblock[0], self)
homelist.append(home)
self.grid.place_agent(home, homeblock[0])
else:
print(
"systemic oppression under capitalism forclosed on one hoomin's home."
)
self.homeset = self.homeset.union(set(homelist))
#initialize hoomins
for i in range(self.initial_hoomins):
x = self.random.randrange(self.width)
y = self.random.randrange(self.height)
hoomin = SocialHoomin(self.next_id(), (x, y), self)
if i == 1:
for x in range(settings.initial_scattermessages):
hoomin.store_scattermessage("hoomin!")
hoomin.pos = (0, 0)
x = 0
y = 0
self.hoomin_zero_id = hoomin.unique_id
if i == self.initial_hoomins - 1:
self.final_hoomin_id = hoomin.unique_id
hoomin.pos = (self.width - 1, self.height - 1)
x = self.width - 1
y = self.height - 1
possiblehomes = self.homeset.difference(Home.claimedhomes)
if len(possiblehomes) > 0:
myhome = self.random.sample(possiblehomes, 1)
if len(myhome) > 0:
myhome[0].claim(hoomin)
self.grid.place_agent(hoomin, (x, y))
self.schedule.add(hoomin)
self.G.add_node(hoomin, agent=[hoomin])
#initialize hoomin friends
friendlist = set(self.schedule._agents)
for i in self.schedule._agents:
fren = self.random.sample(friendlist.difference(set([i])),
settings.friendsperhoomin)
for x in fren:
self.schedule._agents[i].addfriend(x)
self.G.add_edge(self.schedule._agents[i],
self.schedule._agents[x])
self.roadplace_grid()
self.running = True
self.datacollector.collect(self)
def roadplace_grid(self):
for h in range(self.height):
if h % self.gridspacing is 0:
for w in range(self.width):
road = Road(self.next_id(), (w, h), self)
self.grid.place_agent(road, (w, h))
for w in range(self.width):
if w % self.gridspacing is 0:
for h in range(self.height):
road = Road(self.next_id(), (w, h), self)
self.grid.place_agent(road, (w, h))
def roadplace_random(self, direction=0):
if direction is HoominWorld.STRAIGHT:
True
elif direction is HoominWorld.LEFT:
self.roaddir = np.array((-1 * self.roaddir[1], self.roaddir[0]))
elif direction is HoominWorld.RIGHT:
self.roaddir = np.array((self.roaddir[1], -1 * self.roaddir[0]))
else:
self.roaddir = np.array((0, 0))
print("bad bad bad")
# print("placing road, direction ", direction, " coord: ", self.roadcurrentcoord)
newcoord = self.roadcurrentcoord + self.roaddir
if newcoord[0] >= self.width or newcoord[0] < 0:
return None
if newcoord[1] >= self.height or newcoord[1] < 0:
return None
self.roadcurrentcoord += self.roaddir
road = Road(self.next_id(), tuple(self.roadcurrentcoord), self)
self.grid.place_agent(road, tuple(self.roadcurrentcoord))
return road
def singleroad(self, initialcoord=(0, 0)):
#initialize roads
#print("placing road seed: ", initialcoord)
roaddir = self.random.randrange(4)
roadseedx = self.random.randrange(self.width)
roadseedy = self.random.randrange(self.height)
road = Road(self.next_id(), (roadseedx, roadseedy), self)
self.roadcurrentcoord = (roadseedx, roadseedy)
self.grid.place_agent(road, (roadseedx, roadseedy))
#note: roads are not scheduled because they do nothing
road = None
counter = 0
roadlist = []
for i in range(self.initial_roads):
while road is None:
val = self.random.random()
#print("val: " , val)
if val <= self.straightweight:
road = self.roadplace_random(HoominWorld.STRAIGHT)
elif val > self.straightweight and val <= self.leftweight + self.straightweight:
road = self.roadplace_random(HoominWorld.LEFT)
elif val > self.leftweight + self.straightweight:
road = self.roadplace_random(HoominWorld.RIGHT)
if road is None:
#print("err: road is none")
True
roadlist.append(road)
road = None
counter += 1
#print("initialized ", counter, " road tiles")
self.roadset = self.roadset.union(set(roadlist))
del roadlist
def get_hoomin_level(self):
return self.hoomin_level
def logstep(self):
if not self.logger.isopen(HoominWorld.FRIENDNODELOGNAME):
self.logger.open(HoominWorld.FRIENDNODELOGNAME, overwrite=True)
if not self.logger.isopen(HoominWorld.TOTALMESSAGELOGNAME):
self.logger.open(HoominWorld.TOTALMESSAGELOGNAME, overwrite=True)
self.logger.write(
HoominWorld.TOTALMESSAGELOGNAME,
str(self.hoomin_level) + " " + str(self.global_scattermessages))
st = "STEP: " + str(self.hoomin_level) + " hoominzero: " + str(
self.schedule._agents[self.hoomin_zero_id].friendgraph.
number_of_nodes()) + " finalhoomin: " + str(self.schedule._agents[
self.final_hoomin_id].friendgraph.number_of_nodes())
self.logger.write(HoominWorld.FRIENDNODELOGNAME, st)
def step(self):
self.schedule.step()
self.datacollector.collect(self)
self.hoomin_level += 1
if self.hoomin_level % settings.graphrefreshfreq == 0 and settings.displayfriendgraph and not settings.runheadless:
plt.cla()
plt.clf()
nx.draw(self.schedule._agents[self.hoomin_zero_id].friendgraph)
plt.draw()
plt.pause(0.001)
if self.verbose:
print([self.schedule.time, "nothing yet"])
if len(self.schedule._agents[self.final_hoomin_id].scatterbuffer
) >= settings.initial_scattermessages:
print("model completed")
self.running = False
if not self.logger.isopen(HoominWorld.STEPSTOCOMPLETIONLOGNAME):
self.logger.open(HoominWorld.STEPSTOCOMPLETIONLOGNAME,
overwrite=True)
self.logger.write(HoominWorld.STEPSTOCOMPLETIONLOGNAME,
self.hoomin_level)
self.logger.close(HoominWorld.STEPSTOCOMPLETIONLOGNAME)
self.logstep()
def run_model(self, step_count=200):
if self.verbose:
print("Initializing hoomins",
self.schedule.get_hoomin_count(Hoomin))
while self.running:
self.step()
class EmasModel(Model):
description = (
"A model for simulating using EMAS model"
)
def __init__(
self,
height=HEIGHT,
width=WIDTH,
columns=2,
rows=2,
death_level=0,
migration_level=10,
init_energy=100,
moore=True,
reproduction_level=7,
parent_part_to_child=30,
base_child_energy=5,
energy_redistribution_radius=4,
islands=[]
):
super().__init__()
self.height: int = height
self.width: int = width
self.columns: int = columns
self.rows: int = rows
self.death_level: float = death_level
self.migration_level: float = migration_level
self.moore: bool = moore
self.reproduction_level: float = reproduction_level
self.parent_part_to_child: float = parent_part_to_child
self.base_child_energy: float = base_child_energy
self.init_energy: float = init_energy
self.energy_redistribution_radius: int = energy_redistribution_radius
self.islands: List[Tuple[Tuple[int, int], Tuple[int, int]]] = islands
self.grid = MultiGrid(self.height, self.width, torus=True)
borders_x_indexes = sorted([int(self.width * part / columns) for part in range(1, self.columns)])
columns_points: Set[Tuple[int, int]] = {(x, y) for x in borders_x_indexes for y in range(self.height)}
borders_y_indexes = sorted([int(self.height * part / rows) for part in range(1, self.rows)])
rows_points: Set[Tuple[int, int]] = {(x, y) for x in range(self.width) for y in borders_y_indexes}
for border_cords in columns_points | rows_points:
border = IslandBorderAgent(self.next_id(), border_cords, self)
self.grid.place_agent(border, border_cords)
islands_x_corners = [-1] + borders_x_indexes + [self.width]
islands_y_corners = [-1] + borders_y_indexes + [self.height]
# left lower and right upper corner
self.islands = [
((x, y), (islands_x_corners[x_u + 1], islands_y_corners[y_u + 1]))
for x_u, x in enumerate(islands_x_corners) if x != self.width
for y_u, y in enumerate(islands_y_corners) if y != self.height
]
def get_neighborhood(self, pos: Coordinate, include_center=False, radius=1, moore=True):
next_moves = self.grid.get_neighborhood(pos, moore, include_center, radius)
island = self.get_island(pos)
return self._filter_coors_in_island(island, next_moves)
def redistribute_energy(self, pos: Coordinate, energy: float, include_center=False, radius=10):
neighbours = self.grid.get_neighbors(pos, self.moore, include_center, radius)
island = self.get_island(pos)
close_neighbours = list(filter(lambda n: self._is_in_island(island, n.pos), neighbours))
emas_neighbours = list(filter(lambda a: isinstance(a, EmasAgent), close_neighbours))
if emas_neighbours:
energy_delta = energy / len(emas_neighbours)
for neighbour in emas_neighbours:
neighbour.energy += energy_delta
def get_island(self, pos: Coordinate):
return list(filter(lambda coors: coors[0][0] < pos[0] < coors[1][0] and coors[0][1] < pos[1] < coors[1][1],
self.islands)).pop()
def _filter_coors_in_island(self, island: Tuple[Tuple[int, int], Tuple[int, int]],
positions: List[Tuple[int, int]]):
return list(filter(lambda move: island[0][0] < move[0] < island[1][0] and island[0][1] < move[1] <
island[1][1], positions))
def _is_in_island(self, island, pos):
return island[0][0] < pos[0] < island[1][0] and island[0][1] < pos[1] < island[1][1]
def _filter_for_emas_agents(self, agents):
return filter(lambda a: isinstance(a, EmasAgent), agents)
def get_closest_neighbour_on_island(self, pos: Coordinate):
island = self.get_island(pos)
taxi_radius = abs(island[0][0] - island[1][0]) + abs(island[0][1] - island[1][1])
# moore=False -> find neighbourhood using taxi metric
neighbours = self.grid.get_neighbors(pos, False, include_center=False, radius=taxi_radius)
island_neighbours = [
agent for agent in neighbours if self._is_in_island(island, agent.pos) and isinstance(agent, EmasAgent)
]
closest_neighbour = None
closest_distance = taxi_radius
for agent in island_neighbours:
if self.taxi_distance(pos, agent.pos) < closest_distance:
closest_distance = self.taxi_distance(pos, agent.pos)
closest_neighbour = agent
return closest_neighbour
def taxi_distance(self, pos_a: Coordinate, pos_b: Coordinate):
return abs(pos_a[0] - pos_b[0]) + abs(pos_a[1] - pos_b[1])
class ExplorationArea(Model):
def __init__(
self,
nrobots,
wifi_range,
radar_radius=6,
alpha=8.175,
gamma=0.65,
inj_pri=0,
ninjured=None,
ncells=None,
obstacles_dist=None,
load_file=None,
dump_datas=True, # enable data collection
alpha_variation=False, # record datas for alpha variation studies
alpha_csv=alpha_csv, # aggregate datas
alpha_step_csv=alpha_step_csv, # single step datas
gamma_variation=False, # record datas for gamma variation studies
gamma_csv=gamma_csv,
optimization_task=False, # enable a small part of data collection for optimization task
time_csv=number_of_steps_csv,
robot_status_csv=robot_status_csv):
# checking params consistency
if not load_file and (not ncells or not obstacles_dist
or not ninjured):
print("Invalid params")
sys.exit(-1)
# used in server start
self.running = True
self.nrobots = nrobots
self.radar_radius = radar_radius
self.ncells = ncells
self.obstacles_dist = obstacles_dist
self.wifi_range = wifi_range
self.alpha = alpha
self.gamma = gamma
self.ninjured = ninjured
self.inj_pri = inj_pri
self.dump_datas = dump_datas
self.optimization_task = optimization_task
self.frontier = set()
self.broken_beans = 0
# Data collection tools
if self.dump_datas:
# it represents the sum of the difficulties of every cell
self.total_difficulty = 0
self.dc_robot_status = DataCollector({
"idling":
lambda m: self.get_number_robots_status(m, "idling"),
"travelling":
lambda m: self.get_number_robots_status(m, "travelling"),
"exploring":
lambda m: self.get_number_robots_status(m, "exploring"),
"deploying_bean":
lambda m: self.get_number_robots_status(m, "deploying_bean"),
"step":
lambda m: self.get_step(m)
})
self.time_csv = time_csv
self.robot_status_csv = robot_status_csv
if self.optimization_task:
self.total_idling_time = 0
self.alpha_variation = alpha_variation
self.gamma_variation = gamma_variation
if self.alpha_variation:
self.costs_each_path = list()
self.alpha_csv = alpha_csv
self.alpha_step = dict()
self.alpha_step_csv = alpha_step_csv
if self.gamma_variation:
self.gamma_df = pd.DataFrame(columns=["step", "mean", "std"])
self.gamma_csv = gamma_csv
self.schedule = RandomActivation(self)
# unique counter for agents
self.agent_counter = 1
self.nobstacle = 0
# graph of seen cells
self.seen_graph = nx.DiGraph()
rnd.seed()
# place a cell agent for store data and visualization on each cell of the grid
# if map is not taken from file, create it
if load_file == None:
self.grid = MultiGrid(ncells + 2, ncells + 2, torus=False)
for i in self.grid.coord_iter():
if i[1] != 0 and i[2] != 0 and i[1] != self.ncells + 1 and i[
2] != self.ncells + 1:
rand = np.random.random_sample()
obstacle = True if rand < self.obstacles_dist else False
# if obstacle
if obstacle:
self.nobstacle += 1
difficulty = math.inf
explored = -1
priority = 0
utility = -math.inf
# if free
else:
difficulty = np.random.randint(low=1, high=13)
if self.dump_datas:
self.total_difficulty += difficulty
explored = 0
priority = 0
utility = 1.0
# if contour cell
else:
difficulty = np.random.randint(low=1, high=13)
explored = -2
priority = -math.inf
utility = -math.inf
# place the agent in the grid
a = Cell(self.agent_counter, self, i[1:], difficulty, explored,
priority, utility)
self.grid.place_agent(a, i[1:])
self.agent_counter += 1
# create injured agents
valid_coord = []
for i in self.grid.coord_iter():
cell = [
e for e in self.grid.get_cell_list_contents(i[1:])
if isinstance(e, Cell)
][0]
if cell.explored == 0:
valid_coord.append(cell.pos)
for i in range(0, ninjured):
inj_index = rnd.choice(valid_coord)
a = Injured(self.agent_counter, self, inj_index)
self.schedule.add(a)
self.grid.place_agent(a, inj_index)
self.agent_counter += 1
else:
# load map from file
try:
with open(load_file, 'r') as f:
file = f.read()
except:
print("file not found")
sys.exit(-1)
exported_map = literal_eval(file)
self.ncells = int(math.sqrt(len(exported_map["Cell"].keys()))) - 2
self.grid = MultiGrid(self.ncells + 2,
self.ncells + 2,
torus=False)
for index in exported_map["Cell"].keys():
cell = exported_map["Cell"][index]
difficulty = cell[2]
explored = cell[3]
priority = cell[4]
utility = cell[5]
if difficulty == "inf":
difficulty = math.inf
if priority == "-inf":
priority = -math.inf
if utility == "-inf":
utility = -math.inf
if explored == -1:
self.nobstacle += 1
if self.dump_datas and utility == 1:
self.total_difficulty += difficulty
a = Cell(self.agent_counter, self, index, difficulty, explored,
priority, utility)
self.grid.place_agent(a, index)
self.agent_counter += 1
for index in exported_map["Injured"].keys():
a = Injured(self.agent_counter, self, index)
self.schedule.add(a)
self.grid.place_agent(a, index)
self.agent_counter += 1
# create robotic agents
row = 0
starting_coord = []
# data collection number of beans requested
if self.dump_datas:
self.deployed_beans_at_start = 0
# generating the list for the starting position of robots
for c in range(self.grid.width):
# take the agent cell
cell = [
e for e in self.grid.get_cell_list_contents(tuple([c, row]))
if isinstance(e, Cell)
][0]
if cell.explored != -1:
starting_coord.append(c)
for i in range(0, self.nrobots):
column = rnd.choice(starting_coord)
a = Robot(self.agent_counter, self, tuple([column, row]),
self.radar_radius)
self.schedule.add(a)
self.grid.place_agent(a, (column, row))
self.agent_counter += 1
# create initial frontier: add cell in front of the robot if valid and not obstacles
cell = [
e for e in self.grid.get_cell_list_contents(
tuple([column, row + 1])) if isinstance(e, Cell)
][0]
if cell.explored == 0:
self.frontier.add(tuple([column, row + 1]))
try:
cell = [
e for e in self.grid.get_cell_list_contents(
tuple([column + 1, row + 1])) if isinstance(e, Cell)
][0]
if cell.explored == 0:
self.frontier.add(tuple([column + 1, row + 1]))
except:
pass
try:
cell = [
e for e in self.grid.get_cell_list_contents(
tuple([column - 1, row + 1])) if isinstance(e, Cell)
][0]
if cell.explored == 0:
self.frontier.add(tuple([column - 1, row + 1]))
except:
pass
cell = [
e
for e in self.grid.get_cell_list_contents(tuple([column, row]))
if isinstance(e, Cell)
][0]
# in the cell where some robots are deployed, only one bean is deployed
if not cell.wifi_bean:
cell.wifi_bean = True
for index in self.grid.get_neighborhood(
cell.pos,
"moore",
include_center=False,
radius=self.wifi_range):
cell = [
e for e in self.grid.get_cell_list_contents(index)
if isinstance(e, Cell)
][0]
cell.wifi_covered = True
if self.dump_datas:
self.deployed_beans_at_start += 1
# what the model does at each time step
def step(self):
# data collection for alpha variation
if self.alpha_variation:
sim_step = self.get_step(self)
self.alpha_step[sim_step] = list()
# call step function for all of the robots in random order
self.schedule.step()
if self.dump_datas:
self.dc_robot_status.collect(self)
if self.optimization_task:
self.total_idling_time += self.get_number_robots_status(
self, "idling")
if self.gamma_variation:
distances = self.compute_robot_distances(self)
self.gamma_df = self.gamma_df.append(
{
"step": self.get_step(self),
"mean": distances[0],
"std": distances[1]
},
ignore_index=True,
sort=False)
# if all seen cells have benn explored, stop the simulation
# we do this so if there are unreachable cells, the cannot be seen, so the simulation stops anyway
stop_exploration_done = True
for node in self.seen_graph.nodes():
cell = [
obj for obj in self.grid.get_cell_list_contents(node)
if isinstance(obj, Cell)
][0]
if cell.explored == 0 or cell.explored == 1:
stop_exploration_done = False
stop_no_robots = False
if len([x for x in self.schedule.agents if isinstance(x, Robot)]) == 0:
stop_no_robots = True
stop = stop_exploration_done or stop_no_robots
if stop:
print("Simultation ended")
# Data collection
if self.dump_datas:
df = pd.read_csv(self.time_csv)
df = df.append(
{
"nrobots": self.nrobots,
"ncells": self.ncells,
"steps": self.schedule.steps,
"total_difficulty": self.total_difficulty,
"beans_deployed": self.get_number_bean_deployed(self)
},
ignore_index=True)
df.to_csv(self.time_csv, index=False)
df_robots_status = self.dc_robot_status.get_model_vars_dataframe(
)
df = pd.read_csv(self.robot_status_csv)
if len(df["sim_id"]) == 0:
df_robots_status["sim_id"] = 0
else:
df_robots_status["sim_id"] = df["sim_id"][df.index[-1]] + 1
df = df.append(df_robots_status, ignore_index=True, sort=False)
df.to_csv(self.robot_status_csv, index=False)
if self.alpha_variation:
mean = round(np.mean(self.costs_each_path), 3)
std = round(np.std(self.costs_each_path), 3)
df = pd.read_csv(self.alpha_csv)
df = df.append(
{
"nrobots": self.nrobots,
"radar_radius": self.radar_radius,
"alpha": self.alpha,
"gamma": self.gamma,
"mean": mean,
"std": std
},
ignore_index=True)
df.to_csv(self.alpha_csv, index=False)
tmp_df = pd.DataFrame(columns=["step", "cost"])
for s, costs in zip(self.alpha_step.keys(),
self.alpha_step.values()):
if not costs:
tmp_df = tmp_df.append({
"step": s,
"cost": -1
},
ignore_index=True,
sort=False)
continue
for c in costs:
tmp_df = tmp_df.append({
"step": s,
"cost": c
},
ignore_index=True,
sort=False)
df = pd.read_csv(self.alpha_step_csv)
if len(df["sim_id"]) == 0:
tmp_df["sim_id"] = 0
else:
tmp_df["sim_id"] = df["sim_id"][df.index[-1]] + 1
tmp_df["nrobots"] = self.nrobots
tmp_df["radar_radius"] = self.radar_radius
tmp_df["alpha"] = self.alpha
tmp_df["gamma"] = self.gamma
df = df.append(tmp_df, ignore_index=True, sort=False)
df.to_csv(self.alpha_step_csv, index=False)
if self.gamma_variation:
df = pd.read_csv(self.gamma_csv)
if len(df["sim_id"]) == 0:
self.gamma_df["sim_id"] = 0
else:
self.gamma_df["sim_id"] = df["sim_id"][df.index[-1]] + 1
self.gamma_df["nrobots"] = self.nrobots
self.gamma_df["radar_radius"] = self.radar_radius
self.gamma_df["alpha"] = self.alpha
self.gamma_df["gamma"] = self.gamma
df = df.append(self.gamma_df, ignore_index=True, sort=False)
df.to_csv(self.gamma_csv, index=False)
self.running = False
def run_model(self):
while (True):
# search for unexplored cells
stop = True
for node in self.seen_graph.nodes():
cell = [
obj for obj in self.grid.get_cell_list_contents(node)
if isinstance(obj, Cell)
][0]
if cell.explored == 0 or cell.explored == 1:
stop = False
# if all seen cells have benn explored, stop the simulation
# we do this so if there are unreachable cells, the cannot be seen, so the simulation stops anyway
if stop:
self.running = False
break
else:
self.step()
# Data collection utilities
@staticmethod
def get_step(m):
return m.schedule.steps
# these two should go faster since cells are not in the scheduler anymore
@staticmethod
def get_number_robots_status(m, status):
status_value = {
"idling": 0,
"travelling": 1,
"exploring": 2,
"deploying_bean": 3
}
return len([
x for x in m.schedule.agents
if isinstance(x, Robot) and x.status == status_value[status]
])
@staticmethod
def get_number_bean_deployed(m):
return sum([
x.number_bean_deployed
for x in m.schedule.agents if isinstance(x, Robot)
]) + m.deployed_beans_at_start
# function for gamma variation
@staticmethod
def compute_robot_distances(m):
nrobots = m.nrobots
T_up = np.full(
(nrobots, nrobots),
0.0) # if it's only zero, numpy represents only integers
# didn't dig deep in numpy doc but it looks like it handles triangualr matrices as "normal" matrices, so
# i just initilize a full matrix and then i'll use it as a triangular.
robots = [x for x in m.schedule.agents if isinstance(x, Robot)]
# the order of the robots in robots can change from step to step (due to the random scheduler),
# This shouldn't create any type of problem, but to avoid a lot of problems with indexes later on
# we sort them basing on the unique_id
robots.sort(key=lambda x: x.unique_id)
# I need the lowest id to shift back the ids to fit the matrices coordinations
lowest_id = robots[0].unique_id
for r in robots:
matrix_id_row = r.unique_id - lowest_id
# the distance of a robot to itself is zero by definition
y1, x1 = r.pos
for i in range(matrix_id_row + 1, nrobots):
r2 = robots[i] # i can do this because they are sorted
y2, x2 = r2.pos
T_up[matrix_id_row][i] = distance.euclidean([x1, y1], [x2, y2])
mean_dist_robots = list()
# the robot 0 has only the rows
mean_robot_zero = sum(T_up[0, 1:nrobots])
mean_dist_robots.append(mean_robot_zero)
for i in range(1, nrobots -
1): # the last row has no values, i iters the rows
mean_robot = (sum(T_up[0:i, i]) +
sum(T_up[i, i + 1:nrobots])) / (nrobots - 1)
mean_dist_robots.append(mean_robot)
# last robot has only the columns
mean_last_robot = sum(T_up[0:nrobots - 1, nrobots - 1])
mean_dist_robots.append(mean_last_robot)
return tuple([
round(np.mean(mean_dist_robots), 3),
round(np.std(mean_dist_robots), 3)
])
class Model(Model):
def __init__(self, N, B, T, Treg, width, height):
self.num_myelin = N * 8
self.num_agents = N + B + T + Treg + self.num_myelin
self.num_neurons = N
self.num_myelin = 4
self.num_limfocytB = B
self.num_active_B = 0
self.num_infected_B = 0
self.num_limfocytT = T
self.num_activeT = 0
self.num_limfocytTreg = Treg
self.grid = MultiGrid(width, height, True)
self.schedule = RandomActivation(self)
self.available_ids = set()
self.dead_agents = set()
self.new_agents = set()
self.max_id = 0
self.step_count = 1
self.cytokina = 0
self.cytokina_prev = 0
self.sum = 0
self.B = 0
self.a = 0.80
self.Ymax = 100
open('new_and_dead.txt', 'w').close()
# Create agents
neuron_positions = [[3, 3], [3, 10], [3, 20], [3, 27], [10, 3],
[10, 10], [10, 20], [10, 27], [19, 3], [19, 10],
[19, 20], [19, 27], [26, 3], [26, 10], [26, 20],
[26, 27], [14, 15]]
for i in range(self.num_neurons):
a = Neuron(i, self, "Neuron")
self.schedule.add(a)
#Add agent to a random grid cell
#x=self.random.randrange(self.grid.width)
#y=self.random.randrange(self.grid.height)
#self.grid.place_agent(a,(x,y))
pos = neuron_positions[i]
self.grid.place_agent(a, (pos[0], pos[1]))
cells = self.grid.get_neighborhood(a.pos, True, False, 1)
id = self.num_agents - i * 8
for cell in cells:
m = Myelin(id, self, "Myelin")
self.schedule.add(m)
self.grid.place_agent(m, cell)
id -= 1
#dodawanie różnych typów agentów zgodnie z ich liczbą podaną przy inicjacji modelu
for i in range(self.num_limfocytB):
a = LimfocytB(i + self.num_neurons, self, "LimfocytB")
self.schedule.add(a)
#Add agent to a random grid cell
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(a, (x, y))
for i in range(self.num_limfocytT):
a = LimfocytT(i + N + B, self, "LimfocytT")
self.schedule.add(a)
#Add agent to a random grid cell
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(a, (x, y))
for i in range(self.num_limfocytTreg):
a = LimfocytTreg(i + N + B + T, self, "LimfocytTreg")
self.schedule.add(a)
#Add agent to a random grid cell
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(a, (x, y))
self.max_id = self.num_agents - 1
self.datacollector_population = DataCollector(
model_reporters={"Populacja": compute_population})
self.datacollector_T_population = DataCollector(
model_reporters={"Populacja Limfocytów T": T_population})
#self.datacollector_T_precentage=DataCollector(
#model_reporters={"Precentage Limfocyt T": T_popualtion_precentage})
self.datacollector_B_population = DataCollector(
model_reporters={"Populacja Limfocytów B": B_population})
self.datacollector_Treg_population = DataCollector(
model_reporters={"Populacja Limfocytów Treg": Treg_population})
self.datacollector_B_active_population = DataCollector(
model_reporters={
"Populacja Aktywnych Limfocytów B": B_activated_population
})
self.datacollector_T_active_population = DataCollector(
model_reporters={
"Populacja Aktywnych Limfocytów T": T_activated_population
})
self.datacollector_B_infected_population = DataCollector(
model_reporters={
"Populacja Zainfekowanych Limfocytów B": B_infected_population
})
self.datacollector_myelin_population = DataCollector(
model_reporters={
"Populacja osłonek mielinowych": myelin_population
})
self.datacollector_myelin_healths = DataCollector(
model_reporters={
"Suma punktów życia osłonek mielinowych": myelin_healths
})
def step(self):
#print("self running: "+str(self.running()))
self.schedule.step()
self.datacollector_population.collect(self)
self.datacollector_T_population.collect(self)
#self.datacollector_T_precentage.collect(self)
self.datacollector_B_population.collect(self)
self.datacollector_Treg_population.collect(self)
self.datacollector_B_active_population.collect(self)
self.datacollector_T_active_population.collect(self)
self.datacollector_B_infected_population.collect(self)
self.datacollector_myelin_population.collect(self)
self.datacollector_myelin_healths.collect(self)
self.adding_removing()
self.datacollector_myelin_healths.get_model_vars_dataframe().to_csv(
r'Data/myelin_healths25.txt', sep=' ', mode='w')
self.datacollector_T_population.get_model_vars_dataframe().to_csv(
r'Data/T_population25.txt', sep=' ', mode='w')
self.datacollector_B_population.get_model_vars_dataframe().to_csv(
r'Data/B_population25.txt', sep=' ', mode='w')
self.datacollector_Treg_population.get_model_vars_dataframe().to_csv(
r'Data/Treg_population25.txt', sep=' ', mode='w')
self.datacollector_B_active_population.get_model_vars_dataframe(
).to_csv(r'Data/B_active_population25.txt', sep=' ', mode='w')
self.datacollector_T_active_population.get_model_vars_dataframe(
).to_csv(r'Data/T_active_population25.txt', sep=' ', mode='w')
self.datacollector_B_infected_population.get_model_vars_dataframe(
).to_csv(r'Data/B_infected_population25.txt', sep=' ', mode='w')
print("Liczba agentów: " + str(self.num_agents))
print("MaxID: " + str(self.max_id))
self.cytokina = max(
min((self.B + self.cytokina_prev), self.Ymax) * self.a, 0)
print("Cytokina " + str(self.cytokina))
print("Cytokina_prev " + str(self.cytokina_prev))
f = open("agents.txt", 'a')
f.write("======Step : " + str(self.step_count) + "\n")
for agent in self.schedule.agents:
f.write("Agent: " + str(agent.type) + " " + str(agent.unique_id) +
str(agent.pos) + "\n")
f.close()
self.cytokina_prev = self.cytokina
self.B = 0
def running(self):
self.step()
def adding_removing(
self): #funckja odpowiedzialna za dodawanie i usuwanie agentów
#print("AddingRemoving")
f = open("new_and_dead.txt", 'a')
f.write("Step " + str(self.step_count) + "\n")
f.write("======Dead agents======: " + "\n")
for d in self.dead_agents:
try:
self.schedule.remove(d)
self.num_agents -= 1
self.available_ids.add(d.unique_id)
except KeyError:
continue
try:
self.grid._remove_agent(d.pos, d)
except KeyError:
continue
f.write(str(d.unique_id) + " " + d.type + "\n")
#if d.type=="AktywowanyLimfocytT":
# self.cytokina-=1
self.dead_agents.clear()
f.write("======New Agents=====: " + "\n")
for n in self.new_agents:
self.schedule.add(n)
self.num_agents += 1
self.grid.place_agent(n, n.pos)
if n.unique_id in self.available_ids:
self.available_ids.remove(n.unique_id)
f.write(str(n.unique_id) + " " + n.type + "\n")
#if n.type=="AktywowanyLimfocytT":
# self.cytokina+=1
self.new_agents.clear()
m = 1
n = 0
for agent in self.schedule.agents:
if agent.unique_id > m:
m = agent.unique_id
if (agent.type == "LimfocytT") or (agent.type
== "AktywowanyLimfocytT"):
n += 1
self.max_id = m
self.num_limfocytT = 0
self.deficiencies()
f.close()
def deficiencies(self):
n = B_population(self)
if n == 0:
if len(self.available_ids) == 0:
self.max_id += 1
id = self.max_id
else:
id = min(self.available_ids)
self.available_ids.remove(id)
agent = LimfocytB(id, self, "LimfocytB")
self.schedule.add(agent)
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(agent, (x, y))
self.num_agents += 1
n = T_population(self)
if n == 0:
for i in range(10):
if len(self.available_ids) == 0:
self.max_id += 1
id = self.max_id
else:
id = min(self.available_ids)
self.available_ids.remove(id)
agent = LimfocytB(id, self, "LimfocytT")
self.schedule.add(agent)
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(agent, (x, y))
self.num_agents += 1
n = Treg_population(self)
if n == 0:
if len(self.available_ids) == 0:
self.max_id += 1
id = self.max_id
else:
id = min(self.available_ids)
self.available_ids.remove(id)
agent = LimfocytB(id, self, "LimfocytTreg")
self.schedule.add(agent)
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(agent, (x, y))
self.num_agents += 1
class Model(Model):
"""A model with a number of agents.
It represents a space where you can experiment by varying parameters"""
def __init__(self, n_victims, n_offenders, n_criminal_generators, r_criminal_generators, max_cp,
pop_count, width, height):
self.grid = MultiGrid(width, height, torus=False)
self.schedule = RandomActivation(self)
self.datacollector = DataCollector(
model_reporters={"Average Perception of Safety": average_perception_of_safety,
"Crime-rate": crime_rate_single_run},
)
self.width = width
self.height = height
self.running = True
# Number of crimes committed per time step.
self.crime_number = 0.0
# User settable parameters
self.num_victims = n_victims
self.num_offenders = n_offenders
self.num_crime_areas = n_criminal_generators
self.crime_area_rad = r_criminal_generators
self.hotspot_rad = 1
self.max_criminal_preference = max_cp
self.pop_count = pop_count
# add light agents to the grid
self.light_layer()
# add possible victims to the grid
self.add_victims()
# add possible offenders to the grid
self.add_offenders()
# add crime areas
self.add_criminal_areas()
self.datacollector.collect(self)
def light_layer(self):
"""Add the light layer to the model. """
for i in range(self.width):
for j in range(self.height):
l = Light(uuid.uuid4(), i + j, self)
self.schedule.add(l)
# light is ordered from darkest to lightest from the
# bottom left corner to the top right corner.
self.grid.place_agent(l, (i, j))
# light is randomised, so long as the
# luminance difference of surrounding cells is within a limit
# self.grid.place_agent(l, self.r_pos(l))
def r_pos(self, new_l):
"""Randomises the positions of the light realistically by considering the difference
between the illuminance in one position and another"""
# Position precisely one light agent per position.
valid = False
while not valid:
x = self.random.randrange(self.width)
y = self.random.randrange(self.height)
# Check if there is already a light agent in this spot
if not (any(isinstance(agent, Light)
for agent in self.grid.get_cell_list_contents((x, y)))):
# Check whether given the difference in the influencing factors between the agent
# and surrounding agents, the new agent can be placed in this position.
# Returns True or False
if self.valid_pos((x, y), new_l, 1):
valid = True
return x, y
def valid_pos(self, n_pos, new_a, delta):
"""Check whether this position is a valid position given the difference in
the inlfueincing factors between the new agent and the surrounding agents"""
# Get the surrounding positions
neighborhood = self.grid.get_neighborhood(n_pos, moore=True,
include_center=False)
# Get the surrounding positions containing light agents.
cell_contents = []
for pos in neighborhood:
cell_contents.append(self.grid.get_cell_list_contents([pos]))
f_cell_contents = [val for sublist in cell_contents for val in sublist]
surr_light = [a for a in f_cell_contents if isinstance(a, Light)]
# If the surrounding light list is empty, return the position
if len(surr_light) == 0:
return True
else:
# Check whether the difference in illuminance is within the valid scope
for l in surr_light:
if abs(l.illuminance - new_a.illuminance) >= delta:
return False
return True
def get_random_pos(self):
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
return (x, y)
def add_victims(self):
for i in range(self.num_victims):
# Generate a safe location for this agent.
s = SafeLocation(uuid.uuid4(), self)
self.schedule.add(s)
self.grid.place_agent(s, self.get_random_pos())
# Generate the agent given its safe location.
p = PossibleVictim(uuid.uuid4(), self, s)
self.schedule.add(p)
self.grid.place_agent(p, self.get_random_pos())
def add_offenders(self):
for i in range(self.num_offenders):
c = PossibleOffender(uuid.uuid4(), self, self.max_criminal_preference)
self.schedule.add(c)
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(c, (x, y))
def generate_criminal_areas(self):
""" Return a list of criminal areas with specified centroids and radius.
If criminal areas overlap, update the criminal areas.
DOES NOT ADD THE AREAS TO THE MODEL, AS IT IS ONLY THE POSITIONS
WITHIN THE AREAS THAT ARE ADDED TO THE MODEL. """
# ----- HELPER FUNCTION -----
def generate_centroid_list(n):
""" If no centroids have been passed,
generate random centroids for the criminal areas. """
centroids = []
for i in range(n):
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
centroids.append((x, y))
return centroids
def get_surr_pos(pos, r=3):
return self.grid.get_neighborhood(pos, moore=True, include_center=True, radius=r)
# ---------------------------
centroids = self.num_crime_areas
radius = self.crime_area_rad
# If the parameter passed as the centroid is an int, generate this number of centroids.
if isinstance(centroids, int):
centroids = generate_centroid_list(centroids)
criminal_area = []
# For every centroid in the list, create either a hotspot or a criminal area depending
# on the type.
for c in centroids:
criminal_area.append(CrimeGenerator(uuid.uuid4(), self, c, radius))
# Return a list of criminal areas.
return criminal_area
def add_criminal_areas(self, criminal_area=None):
"""Add crime areas to the model. """
# If no criminal area has been passed, generate random criminal area.
if criminal_area is None:
criminal_area = self.generate_criminal_areas()
def increment_crimes(self):
self.crime_number += 1
global crime_number
crime_number += 1
def check_victim_agents(self):
if not [agent for agent in self.schedule.agents if
isinstance(agent, PossibleVictim)]:
self.running = False
def step(self):
"""Advance the model by one step."""
self.schedule.step()
self.datacollector.collect(self)
self.check_victim_agents()
self.crime_number = 0
class SleepAnimals_sinDataColl(Model):
'''
Analysis of the evolution of sleep in animals
'''
# Default values
width = 40
height = 40
number_food_patch = 40
number_sleep_patch = 40
interdistance_factor = 0.7
intradistance_factor = 0.2
fp_depletion_tick = 60
def __init__(self, model_id , genome,width = 40, height = 40,
number_food_patch = 40, number_sleep_patch = 40,
interdistance_factor = 0.7, intradistance_factor = 0.2,
fp_depletion = 60, sleep_and_food_gainfactor = 1):
super().__init__()
# Setting Parameters
self.model_id = model_id
self.width = width
self.height = height
self.number_food_patch = number_food_patch
self.number_sleep_patch = number_sleep_patch
self.interdistance_factor = interdistance_factor
self.intradistance_factor = intradistance_factor
self.sue_factor = sleep_and_food_gainfactor
self.genome = genome
self.fp_center_x = 0
self.fp_center_y = 0
self.sp_center_x = 0
self.sp_center_y = 0
self.current_id_food_patch = 0
self.current_id_sleep_patch = 0
self.fp_tick_to_depletion = fp_depletion
self.schedule = RandomActivation(self)
self.grid = MultiGrid(self.width, self.height, torus = False)
#self.datacollector = DataCollector(
# agent_reporters={"Food_fitness" : lambda a: a.fitness_food,
# "Sleep_fitness" : lambda a: a.fitness_sleep,
# "Fitness" : lambda a: a.fitness,
# "Mode" : lambda a: a.mode,
# "Direction" : lambda a: a.lookingto})
# Picking Centers for Food and Sleep Patches
self.interdistance = 0
self.intradistance = 0
self.findingcenters()
# Populating Food Patches
self.available_food_patches = self.grid.get_neighborhood( (self.fp_center_x, self.fp_center_y), False, True, self.intradistance )
i = 0
while i < self.number_food_patch:
self.new_foodpatch()
i += 1
# Populating Sleep Patches
self.available_sleep_patches = self.grid.get_neighborhood( (self.sp_center_x, self.sp_center_y), False, True, self.intradistance )
i = 0
while i < self.number_sleep_patch:
current_spot = random.choice(self.available_sleep_patches)
if not self.grid.is_cell_empty(current_spot):
continue
sleep_patch = SleepPatch( self.next_id_sleeppatch(), self, current_spot )
self.grid.place_agent( sleep_patch, current_spot )
i += 1
# Adding Animal to the world
current_spot = random.choice( self.grid.empties )
animal = Animal( self.next_id(), self, current_spot, self.genome, self.sue_factor, self.sue_factor )
self.grid.place_agent( animal, current_spot )
self.schedule.add( animal )
def step(self):
'''Advance the model by one step.'''
# self.datacollector.collect(self)
self.schedule.step()
def findingcenters(self):
max_manhattan_length = self.height + self.width
self.interdistance = int(max_manhattan_length * self.interdistance_factor)
self.intradistance = int(max_manhattan_length * self.intradistance_factor)
fp_center_x = 0
fp_center_y = 0
sp_center_x = 0
sp_center_y = 0
centers_selected = False
while not centers_selected:
fp_center_x = random.randrange(self.width)
fp_center_y = random.randrange(self.height)
fp_center = (fp_center_x, fp_center_y)
available_spots_food = self.grid.get_neighborhood(fp_center, False, False, self.intradistance)
if len(available_spots_food) < 80:
continue
if self.interdistance > 0:
list_centers_sp = list( set( self.grid.get_neighborhood(fp_center, False, False, self.interdistance + 1) ) -
set( self.grid.get_neighborhood(fp_center, False, False, self.interdistance - 1) ) )
if len(list_centers_sp) < 5:
continue
else:
list_centers_sp = [(fp_center_x,fp_center_y)]
sp_center = random.choice(list_centers_sp)
(sp_center_x, sp_center_y) = sp_center
available_spots_sleep = self.grid.get_neighborhood(sp_center, False, False, self.intradistance)
if len(available_spots_sleep) < 80:
continue
centers_selected = True
self.fp_center_x = fp_center_x
self.fp_center_y = fp_center_y
self.sp_center_x = sp_center_x
self.sp_center_y = sp_center_y
def next_id_foodpatch(self):
""" Return the next unique ID for food patches, increment current_id"""
self.current_id_food_patch += 1
a = 'M' + str(self.model_id) + 'F' + str(self.current_id_food_patch)
return a
def next_id_sleeppatch(self):
""" Return the next unique ID for sleep patches, increment current_id"""
self.current_id_sleep_patch += 1
a = 'M' + str(self.model_id) + 'S' + str(self.current_id_sleep_patch)
return a
def new_foodpatch(self):
spot_found = False
while not spot_found :
current_spot = random.choice(self.available_food_patches)
if not self.grid.is_cell_empty(current_spot):
continue
food_patch = FoodPatch(self.next_id_foodpatch(), self, current_spot, self.fp_tick_to_depletion)
self.grid.place_agent(food_patch, current_spot)
spot_found = True
def arrayRGB_clusters(self):
available_spots = np.full( (self.grid.width , self.grid.height , 3) , 255)
for coordinates in self.available_sleep_patches:
(x, y) = coordinates
available_spots[x][y]= [100, 0, 150]
for coordinates in self.available_food_patches:
(x, y) = coordinates
available_spots[x][y]= [10, 150, 0]
return available_spots
def arrayRGB_display(self):
RGBdisplay = np.full((self.grid.width, self.grid.height,3), 255)
for cell in self.grid.coord_iter():
cell_content, x, y = cell
a = list(cell_content)
if len(a) == 0:
RGBdisplay[x][y] = [255,255,255]
elif len(a) == 1:
if isinstance(a[0], SleepPatch):
RGBdisplay[x][y] = [100,0,150]
elif isinstance(a[0], FoodPatch):
RGBdisplay[x][y] = [10,150,0]
elif isinstance(a[0], Animal):
RGBdisplay[x][y] = [0,0,0]
else:
pass
elif len(a) == 2:
RGBdisplay[x][y] = [0,0,0]
else:
pass
return RGBdisplay
def array2D_display(self):
display2D = np.zeros((self.grid.width, self.grid.height))
for cell in self.grid.coord_iter():
cell_content, x, y = cell
a = list(cell_content)
if len(a) == 0:
display2D[x][y] = 0
elif len(a) == 1:
if isinstance(a[0], SleepPatch):
display2D[x][y] = 10
elif isinstance(a[0], FoodPatch):
display2D[x][y] = 20
elif isinstance(a[0], Animal):
display2D[x][y] = 30
else:
pass
else:
pass
return display2D
class FireEvacuation(Model):
MIN_HEALTH = 0.75
MAX_HEALTH = 1
MIN_SPEED = 1
MAX_SPEED = 2
MIN_NERVOUSNESS = 1
MAX_NERVOUSNESS = 10
MIN_EXPERIENCE = 1
MAX_EXPERIENCE = 10
MIN_VISION = 1
# MAX_VISION is simply the size of the grid
def __init__(self, floor_plan_file, human_count, collaboration_percentage, fire_probability, visualise_vision, random_spawn, save_plots):
# Load floorplan
# floorplan = np.genfromtxt(path.join("fire_evacuation/floorplans/", floor_plan_file))
with open(os.path.join("fire_evacuation/floorplans/", floor_plan_file), "rt") as f:
floorplan = np.matrix([line.strip().split() for line in f.readlines()])
# Rotate the floorplan so it's interpreted as seen in the text file
floorplan = np.rot90(floorplan, 3)
# Check what dimension our floorplan is
width, height = np.shape(floorplan)
# Init params
self.width = width
self.height = height
self.human_count = human_count
self.collaboration_percentage = collaboration_percentage
self.visualise_vision = visualise_vision
self.fire_probability = fire_probability
self.fire_started = False # Turns to true when a fire has started
self.save_plots = save_plots
# Set up model objects
self.schedule = RandomActivation(self)
self.grid = MultiGrid(height, width, torus=False)
# Used to start a fire at a random furniture location
self.furniture_list = []
# Used to easily see if a location is a FireExit or Door, since this needs to be done a lot
self.fire_exit_list = []
self.door_list = []
# If random spawn is false, spawn_list will contain the list of possible spawn points according to the floorplan
self.random_spawn = random_spawn
self.spawn_list = []
# Load floorplan objects
for (x, y), value in np.ndenumerate(floorplan):
value = str(value)
floor_object = None
if value is "W":
floor_object = Wall((x, y), self)
elif value is "E":
floor_object = FireExit((x, y), self)
self.fire_exit_list.append((x, y))
self.door_list.append((x, y)) # Add fire exits to doors as well, since, well, they are
elif value is "F":
floor_object = Furniture((x, y), self)
self.furniture_list.append((x, y))
elif value is "D":
floor_object = Door((x, y), self)
self.door_list.append((x, y))
elif value is "S":
self.spawn_list.append((x, y))
if floor_object:
self.grid.place_agent(floor_object, (x, y))
self.schedule.add(floor_object)
# Create a graph of traversable routes, used by agents for pathing
self.graph = nx.Graph()
for agents, x, y in self.grid.coord_iter():
pos = (x, y)
# If the location is empty, or a door
if not agents or any(isinstance(agent, Door) for agent in agents):
neighbors = self.grid.get_neighborhood(pos, moore=True, include_center=True, radius=1)
for neighbor in neighbors:
# If there is contents at this location and they are not Doors or FireExits, skip them
if not self.grid.is_cell_empty(neighbor) and neighbor not in self.door_list:
continue
self.graph.add_edge(pos, neighbor)
# Collects statistics from our model run
self.datacollector = DataCollector(
{
"Alive": lambda m: self.count_human_status(m, Human.Status.ALIVE),
"Dead": lambda m: self.count_human_status(m, Human.Status.DEAD),
"Escaped": lambda m: self.count_human_status(m, Human.Status.ESCAPED),
"Incapacitated": lambda m: self.count_human_mobility(m, Human.Mobility.INCAPACITATED),
"Normal": lambda m: self.count_human_mobility(m, Human.Mobility.NORMAL),
"Panic": lambda m: self.count_human_mobility(m, Human.Mobility.PANIC),
"Verbal Collaboration": lambda m: self.count_human_collaboration(m, Human.Action.VERBAL_SUPPORT),
"Physical Collaboration": lambda m: self.count_human_collaboration(m, Human.Action.PHYSICAL_SUPPORT),
"Morale Collaboration": lambda m: self.count_human_collaboration(m, Human.Action.MORALE_SUPPORT)
}
)
# Calculate how many agents will be collaborators
number_collaborators = int(round(self.human_count * (self.collaboration_percentage / 100)))
# Start placing human agents
for i in range(0, self.human_count):
if self.random_spawn: # Place human agents randomly
pos = self.grid.find_empty()
else: # Place human agents at specified spawn locations
pos = random.choice(self.spawn_list)
if pos:
# Create a random human
health = random.randint(self.MIN_HEALTH * 100, self.MAX_HEALTH * 100) / 100
speed = random.randint(self.MIN_SPEED, self.MAX_SPEED)
if number_collaborators > 0:
collaborates = True
number_collaborators -= 1
else:
collaborates = False
# Vision statistics obtained from http://www.who.int/blindness/GLOBALDATAFINALforweb.pdf
vision_distribution = [0.0058, 0.0365, 0.0424, 0.9153]
vision = int(np.random.choice(np.arange(self.MIN_VISION, self.width + 1, (self.width / len(vision_distribution))), p=vision_distribution))
nervousness_distribution = [0.025, 0.025, 0.1, 0.1, 0.1, 0.3, 0.2, 0.1, 0.025, 0.025] # Distribution with slight higher weighting for above median nerovusness
nervousness = int(np.random.choice(range(self.MIN_NERVOUSNESS, self.MAX_NERVOUSNESS + 1), p=nervousness_distribution)) # Random choice starting at 1 and up to and including 10
experience = random.randint(self.MIN_EXPERIENCE, self.MAX_EXPERIENCE)
belief_distribution = [0.9, 0.1] # [Believes, Doesn't Believe]
believes_alarm = np.random.choice([True, False], p=belief_distribution)
human = Human(pos, health=health, speed=speed, vision=vision, collaborates=collaborates, nervousness=nervousness, experience=experience, believes_alarm=believes_alarm, model=self)
self.grid.place_agent(human, pos)
self.schedule.add(human)
else:
print("No tile empty for human placement!")
self.running = True
# Plots line charts of various statistics from a run
def save_figures(self):
DIR = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
OUTPUT_DIR = DIR + "/output"
results = self.datacollector.get_model_vars_dataframe()
dpi = 100
fig, axes = plt.subplots(figsize=(1920 / dpi, 1080 / dpi), dpi=dpi, nrows=1, ncols=3)
status_results = results.loc[:, ['Alive', 'Dead', 'Escaped']]
status_plot = status_results.plot(ax=axes[0])
status_plot.set_title("Human Status")
status_plot.set_xlabel("Simulation Step")
status_plot.set_ylabel("Count")
mobility_results = results.loc[:, ['Incapacitated', 'Normal', 'Panic']]
mobility_plot = mobility_results.plot(ax=axes[1])
mobility_plot.set_title("Human Mobility")
mobility_plot.set_xlabel("Simulation Step")
mobility_plot.set_ylabel("Count")
collaboration_results = results.loc[:, ['Verbal Collaboration', 'Physical Collaboration', 'Morale Collaboration']]
collaboration_plot = collaboration_results.plot(ax=axes[2])
collaboration_plot.set_title("Human Collaboration")
collaboration_plot.set_xlabel("Simulation Step")
collaboration_plot.set_ylabel("Successful Attempts")
collaboration_plot.set_ylim(ymin=0)
timestr = time.strftime("%Y%m%d-%H%M%S")
plt.suptitle("Percentage Collaborating: " + str(self.collaboration_percentage) + "%, Number of Human Agents: " + str(self.human_count), fontsize=16)
plt.savefig(OUTPUT_DIR + "/model_graphs/" + timestr + ".png")
plt.close(fig)
# Starts a fire at a random piece of furniture with file_probability chance
def start_fire(self):
rand = random.random()
if rand < self.fire_probability:
fire_furniture = random.choice(self.furniture_list)
fire = Fire(fire_furniture, self)
self.grid.place_agent(fire, fire_furniture)
self.schedule.add(fire)
self.fire_started = True
print("Fire started at:", fire_furniture)
def step(self):
"""
Advance the model by one step.
"""
self.schedule.step()
# If there's no fire yet, attempt to start one
if not self.fire_started:
self.start_fire()
self.datacollector.collect(self)
# If no more agents are alive, stop the model and collect the results
if self.count_human_status(self, Human.Status.ALIVE) == 0:
self.running = False
if self.save_plots:
self.save_figures()
@staticmethod
def count_human_collaboration(model, collaboration_type):
"""
Helper method to count the number of collaborations performed by Human agents in the model
"""
count = 0
for agent in model.schedule.agents:
if isinstance(agent, Human):
if collaboration_type == Human.Action.VERBAL_SUPPORT:
count += agent.get_verbal_collaboration_count()
elif collaboration_type == Human.Action.MORALE_SUPPORT:
count += agent.get_morale_collaboration_count()
elif collaboration_type == Human.Action.PHYSICAL_SUPPORT:
count += agent.get_physical_collaboration_count()
return count
@staticmethod
def count_human_status(model, status):
"""
Helper method to count the status of Human agents in the model
"""
count = 0
for agent in model.schedule.agents:
if isinstance(agent, Human):
if agent.get_status() == status:
count += 1
return count
@staticmethod
def count_human_mobility(model, mobility):
"""
Helper method to count the mobility of Human agents in the model
"""
count = 0
for agent in model.schedule.agents:
if isinstance(agent, Human):
if agent.get_mobility() == mobility:
count += 1
return count
class InfectionModel(Model):
"""A model for infection spread."""
def __init__(self,
human_count,
random_spawn,
save_plots,
floor_plan_file="floorplan_1.txt",
N=30,
width=10,
height=10,
ptrans=0.5,
progression_period=3,
progression_sd=2,
death_rate=0.0193,
recovery_days=21,
recovery_sd=7):
# Load floorplan
# floorplan = np.genfromtxt(path.join("infection_spread/floorplans/", floor_plan_file))
with open(
os.path.join("infection_spread/floorplans/", floor_plan_file),
"rt") as f:
floorplan = np.matrix(
[line.strip().split() for line in f.readlines()])
# Rotate the floorplan so it's interpreted as seen in the text file
floorplan = np.rot90(floorplan, 3)
# Check what dimension our floorplan is
width, height = np.shape(floorplan)
# Init params
self.num_agents = N
self.initial_outbreak_size = 1
self.recovery_days = recovery_days
self.recovery_sd = recovery_sd
self.ptrans = ptrans
self.death_rate = death_rate
self.running = True
self.dead_agents = []
self.width = width
self.height = height
self.human_count = human_count
# Set up model objects
self.schedule = RandomActivation(self)
self.grid = MultiGrid(height, width, torus=False)
# Used to easily see if a location is an Exit or Door, since this needs to be done a lot
self.exit_list = []
self.door_list = []
# If random spawn is false, spawn_list will contain the list of possible spawn points according to the floorplan
self.random_spawn = random_spawn
self.spawn_list = []
# Load floorplan objects
for (x, y), value in np.ndenumerate(floorplan):
value = str(value)
floor_object = None
if value == "W":
floor_object = Wall((x, y), self)
elif value == "E":
floor_object = Exit((x, y), self)
self.exit_list.append((x, y))
self.door_list.append((x, y)) # Add exits to doors as well
elif value == "D":
floor_object = Door((x, y), self)
self.door_list.append((x, y))
elif value == "S":
self.spawn_list.append((x, y))
if floor_object:
self.grid.place_agent(floor_object, (x, y))
self.schedule.add(floor_object)
# Create a graph of traversable routes, used by agents for pathing
self.graph = nx.Graph()
for agents, x, y in self.grid.coord_iter():
pos = (x, y)
# If the location is empty, or a door
if not agents or any(isinstance(agent, Door) for agent in agents):
neighbors = self.grid.get_neighborhood(pos,
moore=True,
include_center=True,
radius=1)
for neighbor in neighbors:
# If there is contents at this location and they are not Doors or Exits, skip them
if not self.grid.is_cell_empty(
neighbor) and neighbor not in self.door_list:
continue
self.graph.add_edge(pos, neighbor)
# Collects statistics from our model run
self.datacollector = DataCollector({
"Susceptible":
lambda m: self.count_human_status(m, Human.State.SUSCEPTIBLE),
"Infected":
lambda m: self.count_human_status(m, Human.State.INFECTED),
"Removed":
lambda m: self.count_human_status(m, Human.State.REMOVED)
})
# Start placing human agents
# for i in range(0, self.human_count):
# if self.random_spawn: # Place human agents randomly
# pos = self.grid.find_empty()
# else: # Place human agents at specified spawn locations
# pos = random.choice(self.spawn_list)
#if pos:
# Create a random human
# health = random.randint(self.MIN_HEALTH * 100, self.MAX_HEALTH * 100) / 100
# speed = random.randint(self.MIN_SPEED, self.MAX_SPEED)
for i in range(0, self.num_agents):
a = Human(unique_id=i, model=self)
self.schedule.add(a)
# Add the agent to a random grid cell
# x = self.random.randrange(self.grid.width)
# y = self.random.randrange(self.grid.height)
# self.grid.place_agent(a, (x, y))
if self.random_spawn: # Place human agents randomly
pos = self.grid.find_empty()
else: # Place human agents at specified spawn locations
pos = random.choice(self.spawn_list)
self.grid.place_agent(a, pos)
#make some agents infected at start
infected = np.random.choice([0, 1], 1, p=[0.98, 0.02])
if infected == 1:
a.state = State.INFECTED
a.recovery_time = self.get_recovery_time()
self.datacollector = DataCollector(
#model_reporters={"Gini": compute_gini},
agent_reporters={"State": "state"})
def get_recovery_time(self):
return int(
self.random.normalvariate(self.recovery_days, self.recovery_sd))
def step(self):
"""
Advance the model by one step.
"""
self.schedule.step()
self.datacollector.collect(self)
class CancerModel(Model):
def __init__(self,cancer_cells_number,cure_number,eat_values, verovatnoca_mutacije):
self.counter = 0
self.cure_number = cure_number
radoznalosti = list(np.arange(0.01,KILLING_PROBABILITY,0.01))
print(radoznalosti)
self.datacollector = DataCollector(
model_reporters = {"FitnessFunction":fitness_funkcija,
"SpeedSum":overall_speed,"SmartMedicine":num_of_smart_medicine,
"RadoznalostSum":radoznalost_sum })
grid_size = math.ceil(math.sqrt(cancer_cells_number*4))
self.grid = MultiGrid(grid_size,grid_size,False)
speeds = list(range(grid_size//2)) #popravi te boje TODO
print(speeds)
poss = self.generate_cancer_cell_positions(grid_size,cancer_cells_number)
num_CSC = math.ceil(percentage(1,cancer_cells_number))
pos_CSC = [self.random.choice(poss) for i in range(num_CSC)]
self.schedule = RandomActivation(self)
self.running = True
for i in range(cancer_cells_number):
pos = poss[i]
c = CancerStemCell(uuid.uuid4(),self,value = eat_values[CancerStemCell.__class__]) if pos in pos_CSC else CancerCell(i,self,value=eat_values[CancerCell.__class__])
self.grid.place_agent(c,pos)
self.schedule.add(c)
for i in range(cure_number):
#pos = self.grid.find_empty()
pos = (0,0)
radoznalost = self.random.choice(radoznalosti)
speed = self.random.choice(speeds)
a = CureAgent(uuid.uuid4(),self,speed = speed,radoznalost=radoznalost) if i< cure_number//2 else SmartCureAgent(uuid.uuid4(),self,speed=speed,radoznalost = radoznalost)
self.grid.place_agent(a,pos)
self.schedule.add(a)
for (i,(contents, x,y)) in enumerate(self.grid.coord_iter()):
if not contents:
c = HealthyCell(uuid.uuid4(),self,eat_values[HealthyCell.__class__])
self.grid.place_agent(c,(x,y))
self.schedule.add(c)
def generate_cancer_cell_positions(self,grid_size,cancer_cells_number):
center = grid_size//2
poss = [(center,center)]
for pos in poss:
poss+=[n for n in self.grid.get_neighborhood(pos,moore=True,include_center=False) if n not in poss]
if len(set(poss))>=cancer_cells_number:
break
poss = list(set(poss))
return poss
def duplicate_or_kill(self):
koliko = math.ceil(percentage(5,self.cure_number)) # TODO igor javlja kako biramo procena
cureagents = [c for c in self.schedule.agents if isinstance(c,CureAgent)]
sortirani = sorted(cureagents, key=lambda x: x.points, reverse=True)
poslednji = sortirani[-koliko:]
prvi = sortirani[:koliko]
assert(len(prvi)==len(poslednji))
self.remove_agents(poslednji)
self.duplicate_agents(prvi)
def remove_agents(self,agents):
for a in agents:
self.schedule.remove(a)
self.grid.remove_agent(a)
def duplicate_agents(self,agents):
for a in agents:
a_new = a.__class__(uuid.uuid4(),model=self,speed = a.speed,radoznalost = a.radoznalost) #TODO ostale parametre isto poistoveti
self.grid.place_agent(a_new,(1,1))
self.schedule.add(a_new)
def step(self):
self.datacollector.collect(self)
self.counter+=1
self.schedule.step()
if self.counter%10 ==0: # TODO ovo menjamo, parameter TODO
#TODO sredi boje i
#TODO sredi ovo pucanje zbog nule u latin hypercube
#TODO napravi da je R promenljivo
self.duplicate_or_kill()
class PandemicsModel(Model):
def __init__(self, config=default_config, disease=dis.covid_disease):
self.agents_count = config.citizens_count + config.policemen_count
self.disease = disease
self.deceased = []
self.buried = []
self.deceased_counter = 0
self.infected_counter = 0
self.grid = MultiGrid(config.width, config.height, True)
self.safety_per_cell = np.ones((config.height, config.width))
self.buildings_map = np.zeros((config.height, config.width))
self.buildings_id_map = np.zeros((config.height, config.width))
self.schedule = SimultaneousActivation(self)
self.datacollector = DataCollector(
model_reporters={
"deceased": "deceased_counter",
"infected": "infected_counter"},
agent_reporters={
"hp": lambda a: a.profile["hp"],
"mask_protection": "mask_protection",
"infection_day": lambda a: a.profile["infection_day"],
"obedience": lambda a: a.profile["obedience"],
"fear": lambda a: a.profile["fear"]}
)
self.config = config
self.buildings = {b["id"] : b for b in self.config.buildings}
self.houses = [x for x in self.buildings.values() if x['type'] == 'house']
self.workplaces = [x for x in self.buildings.values() if x['type'] == 'workplace']
self.shops = [x for x in self.buildings.values() if x['type'] == 'shop']
self.add_buildings_to_map(self.buildings)
self.street_positions = []
for x in range(self.config.width):
for y in range(self.config.height):
if self.buildings_map[y][x] == 0:
self.street_positions.append((x, y))
self.house_to_agents = defaultdict(list)
self.workplace_to_agents = defaultdict(list)
self.current_location_type = None
# Create agents
for i in range(self.agents_count):
if i < config.policemen_count:
a = agent.create_distribution_policeman_agent(
i, self, config.policemen_mental_features_distribution)
a.assign_house(self, self.houses)
elif i < config.policemen_count + config.citizens_count:
a = agent.create_distribution_citizen_agent(
i, self, config.citizens_mental_features_distribution)
a.assign_house(self, self.houses)
a.assign_workplace(self, self.workplaces)
self.add_agent(a)
for i in self.random.choices(self.schedule.agents, k=config.infected_count):
i.start_infection()
self.running = True
self.steps_count = 0
self.datacollector.collect(self)
# Returns (type, id) of the building where agent a is currently located
def where_is_agent(self, a):
(x, y) = a.pos
return (self.buildings_map[y][x], self.buildings_id_map[y][x])
def compute_time_of_day(self):
return self.steps_count % self.config.steps_per_day / (self.config.steps_per_day / HOURS_PER_DAY)
def compute_current_location_type(self):
t = self.compute_time_of_day()
return self.config.day_plan[t] if t in self.config.day_plan else \
self.config.day_plan[min(self.config.day_plan.keys(), key=lambda k: k-t)]
# Updates current location type based on time of day and the config schedule
# Returns true if there is a change in current_location_type
def update_current_location_type(self):
t = self.compute_time_of_day()
if t in self.config.day_plan:
self.current_location_type = self.config.day_plan[t]
return True
return False
def add_buildings_to_map(self, buildings):
for b in buildings.values():
(x, y) = b["bottom-left"]
for i in range(x, x+b["width"]):
for j in range(y, y+b["height"]):
self.buildings_map[j][i] = self.config.building_tags[b['type']]
self.buildings_id_map[j][i] = b['id']
def add_agent(self, a):
self.schedule.add(a)
# Add the agent to a random grid cell
x = self.random.randrange(self.grid.width)
y = self.random.randrange(self.grid.height)
self.grid.place_agent(a, (x, y))
def bury_agent(self, a):
self.schedule.remove(a)
self.grid.remove_agent(a)
self.deceased_counter += 1
self.buried.append(a)
def risk_from_agents(self, agents, weight):
risk = 0
for a in agents:
risk += 1 - a.mask_protection
return risk*weight
def evaluate_safety_per_cell(self):
""" 1.0 is a perfectly safe cell - empty, with all neighbours
and neighbours-of-neighbours empty as well """
for content, x, y in model.grid.coord_iter():
self.safety_per_cell[y][x] = 1 # initial value
# Compute risk from (x,y) cell
ring0_risk = self.risk_from_agents(content, weight=0.5)
considered_cells = {(x,y)}
# Compute risk coming from the neighbours of (x,y) cell
neighbours = self.grid.get_neighborhood(
(x,y), moore=True, include_center=False)
neighbours_content = self.grid.get_cell_list_contents(neighbours)
ring1_risk = self.risk_from_agents(neighbours_content, 0.25)
considered_cells | set(neighbours)
# Compute risk coming from
# the neighbours of the neighbours of (x,y) cell
neighbours_of_neighbours = set()
for c in neighbours:
neighbours_of_neighbours | set(
self.grid.get_neighborhood(
(x,y),moore=True, include_center=False))
neighbours_of_neighbours -= considered_cells
ring2_risk = self.risk_from_agents(
self.grid.get_cell_list_contents(neighbours_of_neighbours), 0.125)
self.safety_per_cell[y][x] -= ring0_risk + ring1_risk + ring2_risk
def step(self):
if (self.update_current_location_type()):
for a in self.schedule.agents:
if (self.current_location_type is not None):
b = a.select_building(self.current_location_type)
a.teleport_to_building(b)
else:
a.teleport_to_street()
self.evaluate_safety_per_cell()
self.schedule.step()
self.steps_count += 1
# collect data
self.datacollector.collect(self)
for d in self.deceased:
self.bury_agent(d)
self.deceased = []
def run_model(self, n):
for i in range(n):
self.step()
class Environment(Model):
""" A model which contains a number of ant colonies. """
def __init__(self, width, height, n_colonies, n_ants, n_obstacles, decay=0.2, sigma=0.1, moore=False, birth=True, death=True):
"""
:param width: int, width of the system
:param height: int, height of the system
:param n_colonies: int, number of colonies
:param n_ants: int, number of ants per colony
:param decay: float, the rate in which the pheromone decays
:param sigma: float, sigma of the Gaussian convolution
:param moore: boolean, True/False whether Moore/vonNeumann is used
"""
super().__init__()
# Agent variables
self.birth = birth
self.death = death
self.pheromone_level = 1
# Environment variables
self.width = width
self.height = height
self.grid = MultiGrid(width, height, False)
self.moore = moore
self.sigma = sigma
self.decay = decay
# Environment attributes
self.schedule = RandomActivation(self)
self.colonies = [Colony(self, i, (width // 2, height // 2), n_ants, birth=self.birth, death=self.death) for i in range(n_colonies)]
self.pheromones = np.zeros((width, height), dtype=np.float)
self.pheromone_updates = []
self.food = FoodGrid(self)
self.food.add_food()
self.obstacles = []
for _ in range(n_obstacles):
self.obstacles.append(Obstacle(self))
# Metric + data collection
self.min_distance = distance.cityblock(self.colonies[0].pos, self.food.get_food_pos())
self.datacollector = DataCollector(
model_reporters={"Minimum path length": metrics.min_path_length,
"Mean minimum path length": metrics.mean_min_path_length},
agent_reporters={"Agent minimum path length": lambda x: min(x.path_lengths),
"Encounters": Ant.count_encounters})
# Animation attributes
self.pheromone_im = None
self.ax = None
def step(self):
"""
Do a single time-step using freeze-dry states, colonies are updated each time-step in random orders, and ants
are updated per colony in random order.
"""
self.food.step()
self.datacollector.collect(self)
# Update all colonies
for col in random.sample(self.colonies, len(self.colonies)):
col.step()
self.schedule.step()
self.update_pheromones()
def move_agent(self, ant, pos):
"""
Move an agent across the map.
:param ant: class Ant
:param pos: tuple (x, y)
"""
if self.moore:
assert np.sum(np.subtract(pos, ant.pos) ** 2) in [1, 2], \
"the ant can't move from its original position {} to the new position {}, because the distance " \
"is too large".format(ant.pos, pos)
else:
assert np.sum(np.subtract(pos, ant.pos) ** 2) == 1, \
"the ant can't move from its original position {} to the new position {}, because the distance " \
"is too large, loc_food {}".format(ant.pos, pos, self.food.get_food_pos())
self.grid.move_agent(ant, pos)
def get_random_position(self):
return (np.random.randint(0, self.width), np.random.randint(0, self.height))
def position_taken(self, pos):
if pos in self.food.get_food_pos():
return True
for colony in self.colonies:
if colony.on_colony(pos):
return True
for obstacle in self.obstacles:
if obstacle.on_obstacle(pos):
return True
return False
def add_food(self):
"""
Add food somewhere on the map, which is not occupied by a colony yet
"""
self.food.add_food()
def place_pheromones(self, pos):
"""
Add pheromone somewhere on the map
:param pos: tuple (x, y)
"""
self.pheromone_updates.append((pos, self.pheromone_level))
def get_neighbor_pheromones(self, pos, id):
"""
Get the passable neighboring positions and their respective pheromone levels for the pheromone id
:param pos:
:param id:
:return:
"""
indices = self.grid.get_neighborhood(pos, self.moore)
indices = [x for x in indices if not any([isinstance(x, Obstacle) for x in self.grid[x[0]][x[1]]])]
pheromones = [self.pheromones[x, y] for x, y in indices]
return indices, pheromones
def update_pheromones(self):
"""
Place the pheromones at the end of a timestep on the grid. This is necessary for freeze-dry time-steps
"""
for (pos, level) in self.pheromone_updates:
# self.pheromones[pos] += level
self.pheromones[pos] += 1
self.pheromone_updates = []
# gaussian convolution using self.sigma
self.pheromones = gaussian_filter(self.pheromones, self.sigma) * self.decay
def animate(self, ax):
"""
:param ax:
:return:
"""
self.ax = ax
self.animate_pheromones()
self.animate_colonies()
self.animate_ants()
self.animate_food()
self.animate_obstacles()
def animate_pheromones(self):
"""
Update the visualization part of the Pheromones.
:param ax:
"""
pheromones = np.rot90(self.pheromones.astype(np.float64).reshape(self.width, self.height))
if not self.pheromone_im:
self.pheromone_im = self.ax.imshow(pheromones,
vmin=0, vmax=50,
interpolation='None', cmap="Purples")
else:
self.pheromone_im.set_array(pheromones)
def animate_colonies(self):
"""
Update the visualization part of the Colonies.
:return:
"""
for colony in self.colonies:
colony.update_vis()
def animate_food(self):
"""
Update the visualization part of the FoodGrid.
:return:
"""
self.food.update_vis()
def animate_ants(self):
"""
Update the visualization part of the Ants.
"""
for ant in self.schedule.agents:
ant.update_vis()
def animate_obstacles(self):
"""
Update the visualization part of the Obstacles.
:return:
"""
for obstacle in self.obstacles:
obstacle.update_vis()
def grid_to_array(self, pos):
"""
Convert the position/indices on self.grid to imshow array.
:param pos: tuple (int: x, int: y)
:return: tuple (float: x, float: y)
"""
return pos[0] - 0.5, self.height - pos[1] - 1.5
def pheromone_threshold(self, threshold):
""" Returns an array of the positions in the grid in which
the pheromone levels are above the given threshold"""
pher_above_thres = np.where(self.pheromones >= threshold)
return list(zip(pher_above_thres[0],pher_above_thres[1]))
def find_path(self, pher_above_thres):
""" Returns the shortest paths from all the colonies to all the food sources.
A path can only use the positions in the given array. Therefore, this function
checks whether there is a possible path for a certain pheremone level.
Essentially a breadth first search"""
space_searched = False
all_paths = []
food_sources = self.food.get_food_pos()
# Search the paths for a colony to all food sources
for colony in self.colonies:
colony_paths = []
pos_list = {colony.pos} # Prooning
possible_paths = [[colony.pos]]
# Continue expanding search area until all food sources found
# or until the entire space is searched
while food_sources != [] and not space_searched:
space_searched = True
temp = []
for path in possible_paths:
for neighbor in self.grid.get_neighborhood(include_center=False, radius=1, pos=path[-1], moore=self.moore):
if neighbor in food_sources:
food_path = copy(path)
food_path.append(neighbor)
colony_paths.append(food_path)
food_sources.remove(neighbor)
# Add epanded paths to the possible paths
if neighbor in pher_above_thres and neighbor not in pos_list:
space_searched = False
temp_path = copy(path)
temp_path.append(neighbor)
temp.append(temp_path)
pos_list.add(neighbor)
possible_paths.remove(path)
possible_paths += temp
all_paths.append(colony_paths)
return all_paths
class World(Model):
def __init__(self, N, coop, width=100, height=100):
self.num_agents = N
self.grid = MultiGrid(width, height, False)
self.schedule = RandomActivation(self)
self.running = True
self.population_center_x = np.random.randint(
POP_MARGIN, self.grid.width - POP_MARGIN)
self.population_center_y = np.random.randint(
POP_MARGIN, self.grid.height - POP_MARGIN)
self.num_energy_resources = RESERVE_SIZE
self.ages = []
self.expected_ages = []
self.num_explorers = 0
self.num_exploiters = 0
self.datacollector = DataCollector(model_reporters={
"Num_Explorer": "num_explorers",
"Num_Exploiter": "num_exploiters"
})
self.bases = self.init_base()
# add social agents to the world
for i in range(1, self.num_agents + 1):
if np.random.uniform() <= EXPLORER_RATIO:
# create a new explorer
a = Explorer("explorer_%d" % (i), self, coop)
self.num_explorers += 1
else:
# create a new exploiter
a = Exploiter("exploiter_%d" % (i), self, coop)
self.num_exploiters += 1
# keep society members confined at beginning
x = np.random.randint(self.population_center_x - POP_SPREAD,
self.population_center_x + POP_SPREAD)
y = np.random.randint(self.population_center_y - POP_SPREAD,
self.population_center_y + POP_SPREAD)
self.grid.place_agent(a, (x, y))
# add agent to scheduler
self.schedule.add(a)
self.expected_ages.append(a.energy / a.living_cost)
# add energy reserves to the world
for i in range(self.num_energy_resources):
a = EnergyResource("energy_reserve_%d" % (i), self)
# decide location of energy reserve
x = np.random.randint(0, self.grid.width)
y = np.random.randint(0, self.grid.height)
self.grid.place_agent(a, (x, y))
self.schedule.add(a)
def init_base(self):
pos = (self.population_center_x, self.population_center_y)
return self.grid.get_neighborhood(pos, True, radius=POP_SPREAD)
def step(self):
self.datacollector.collect(self)
self.schedule.step()
if self.schedule.steps % NEW_ENERGY_STEPS == 0:
num_dead = len(self.ages)
remaining = self.num_agents - num_dead
if np.random.uniform() < NEW_ENERGY_PROB and remaining > 10:
self.num_energy_resources += 1
a = EnergyResource(
"energy_reserve_%d" % (self.num_energy_resources), self)
# decide location of energy reserve
x = np.random.randint(0, self.grid.width)
y = np.random.randint(0, self.grid.height)
self.grid.place_agent(a, (x, y))
self.schedule.add(a)
class CancerModel(Model):
def xprint(self, *args):
logger.info("CANCER MODEL: " + " ".join(map(str, args)))
def __init__(self, cure_agent_type, config):
self.xprint("STARTING SIMULATION !!!")
self.counter = 0
self.decay_number = 0
self.lifetime_counter = 0
self.metastasis_score = 0
eat_values = {CancerCell: 1, HealthyCell: -1, CancerStemCell: 5}
assert (issubclass(cure_agent_type, CureAgent))
agent_memory_range = config["Agent"]["memory_limit"]
agent_memory_type = config["Agent"]["memory_type"]
radoznalost = config["Agent"]["curiosity"]
turn_off_modifiers = config["Tumor"]["turn_off_modifiers"]
CC_mutation_probability = config["Tumor"]["mutation_probability"]
is_tumor_growing = config["Tumor"]["is_growing"]
tumor_movement_stopping_range = config["Tumor"][
"movement_stopping_range"]
steps_before_mutation = config["Model"]["steps_before_mutation"]
self.SAMPLE_i = config["Model"]["sample_i"]
self.cure_number = config["Model"]["NA_number"]
self.probabilites_ranges = config["Agent"]["probabilities_limits"]
self.modifier_fraction = config["Tumor"]["modifier_fraction"]
self.mode = config["Simulation"]["mode"]
fname = "xxx" if self.mode == "learning" else config["Simulation"][
"fname"]
self.MUTATION_PERCENTAGE = config["Model"]["mutation_percentage"]
tumor_growth_probability = config["Tumor"]["growth_probability"]
cancer_cell_number = config["Model"]["CC_number"]
#DATA COLLECTION
self.datacollector = DataCollector(
model_reporters={
"FitnessFunction": fitness_funkcija,
"AverageSpeed": speed_avg,
"AverageMemoryCapacity": memory_size_all_avg,
"PopulationHeterogenity": population_heterogenity,
"MutationAmount": mutation_amount,
"CancerStemCell Number": CSC_number,
"CSC Specialized Agents": CSC_specialized_agents,
"CancerHeterogenity1": cancer_heterogenity_1,
"CancerHeterogenity2": cancer_heterogenity_2,
"CC_Number": CC_number,
"HealthyCell_Number": HC_number,
"MetastasisScore": "metastasis_score",
"CancerSize": cancer_size,
"TotalTumorResiliance": overall_cancer_resiliance,
"TumorResiliance_Pi": cancer_resiliance_Pi,
"TumorResiliance_Pd": cancer_resiliance_Pd,
"TumorResiliance_Pa": cancer_resiliance_Pa,
"TumorResiliance_Pk": cancer_resiliance_Pk,
"TumorResiliance_Psd": cancer_resiliance_Psd,
"NumberOfMutatedCells": mutated_CCs_num,
"TumorCoverage": tumor_coverage,
"AveragePd": average_Pd,
"AveragePa": average_Pa,
"AveragePi": average_Pi,
"AveragePk": average_Pk,
"AveragePsd": average_Psd,
# "PopulationClusters":cluster_counts
},
agent_reporters={
"Pi": get_Pi,
"Pa": get_Pa,
"Pd": get_Pd,
"speed": get_speed,
"Psd": get_Psd,
"Pk": get_Pk,
"memory_size": get_memory_size,
"type": get_agent_type
})
grid_size = math.ceil(math.sqrt(cancer_cell_number * 4))
self.STEPS_BEFORE_MUTATION = steps_before_mutation
self.grid = MultiGrid(grid_size, grid_size, False)
self.NUM_OF_INJECTION_POINTS = config["Model"]["injection_points"]
# self.speeds = list(range(1,grid_size//2))
self.speeds = [
1
] #TODO ovo mozda bolje? ne znam da li treba u config fajlu?
poss = self.generate_cancer_cell_positions(grid_size,
cancer_cell_number)
num_CSC = math.ceil(percentage(1, cancer_cell_number))
pos_CSC = [self.random.choice(poss) for i in range(num_CSC)]
#ACTIVATE SIMULATION
self.schedule = RandomActivation(self)
self.running = True
#PLACE CANCER CELLS
for i in range(cancer_cell_number):
pos = poss[i]
has_modifiers = False if ((i < (
(1 - self.modifier_fraction) * cancer_cell_number))
or turn_off_modifiers is True
) else True #10 % will be with modifiers
c = CancerStemCell("CANCER_STEM_CELL-"+str(uuid.uuid4()),self,value = eat_values[CancerStemCell],has_modifiers=has_modifiers,mutation_probability=CC_mutation_probability,grows=is_tumor_growing,growth_probability=tumor_growth_probability) \
if pos in pos_CSC else CancerCell("CANCER_CELL-"+str(uuid.uuid4()),self,value=eat_values[CancerCell],has_modifiers=has_modifiers,mutation_probability=CC_mutation_probability,grows=is_tumor_growing,growth_probability=tumor_growth_probability)
self.grid.place_agent(c, pos)
self.schedule.add(c)
#PLACE HEALTHY CELLS
for (i, (contents, x, y)) in enumerate(self.grid.coord_iter()):
nbrs = self.grid.get_neighborhood([x, y], moore=True)
second_nbrs = []
for nbr in nbrs:
second_nbrs += self.grid.get_neighborhood(nbr, moore=True)
nbrs = nbrs + second_nbrs
nbr_contents = self.grid.get_cell_list_contents(nbrs)
nbr_CCs = [
nbr for nbr in nbr_contents if isinstance(nbr, CancerCell)
]
if not contents and len(nbr_CCs) == 0:
c = HealthyCell(uuid.uuid4(), self, eat_values[HealthyCell])
self.grid.place_agent(c, (x, y))
self.schedule.add(c)
if self.mode == "simulation":
self.duplicate_mutate_or_kill = self.simulation_mode_function
self.decay_probability = 0.04 #MAGIC NUMBER
self.read_nanoagents_from_file(
fname=fname,
cure_agent_type=cure_agent_type,
tumor_movement_stopping_range=tumor_movement_stopping_range,
agent_memory_range=agent_memory_range,
radoznalost=radoznalost)
elif self.mode == "learning":
self.decay_probability = 0
self.make_nanoagents_from_scratch(cure_agent_type, radoznalost,
agent_memory_type,
agent_memory_range,
tumor_movement_stopping_range,
grid_size)
else:
assert ("False")
def get_random_positions_on_empty_cells(self):
"""Gets the N random currently empty positions on the grid"""
#GETTING EMPTY POSITIONS (for placing nano agents)
empty_cells = [(x, y) for (i, (contents, x,
y)) in enumerate(self.grid.coord_iter())
if not contents]
positions = [
self.random.choice(empty_cells)
for i in range(self.NUM_OF_INJECTION_POINTS)
]
self.xprint("The 5 random empty positions are %s" % positions)
return positions
def inject_nanoagents(self):
"""Injects the nanoagents, they will be activated slowly after"""
from itertools import cycle
positions = cycle(self.get_random_positions_on_empty_cells())
for a in self.agents:
self.grid.place_agent(a, next(positions))
self.agents_iterator = iter(self.agents)
def activate_next_batch_of_agents(self):
agents_to_be_placed_at_each_steps = round(len(self.agents) /
14) #MAGIC NUMBER
for i in range(agents_to_be_placed_at_each_steps):
self.schedule.add(next(self.agents_iterator))
def read_nanoagents_from_file(self, fname, cure_agent_type, radoznalost,
agent_memory_range,
tumor_movement_stopping_range):
import pandas as pd
self.xprint("Reading nanoagents from file")
df = pd.read_csv(fname)
self.agents = []
for i, row in df.iterrows():
a = cure_agent_type(
uuid.uuid4(),
self,
speeds=self.speeds,
radoznalost=radoznalost,
memory_type=row.memory_size,
memory_range=agent_memory_range,
tumor_movement_stopping_range=tumor_movement_stopping_range,
probabilities_ranges=self.probabilites_ranges)
a.Pi = row.Pi
a.Pa = row.Pa
a.Pd = row.Pd
a.memory_size = row.memory_size
a.memorija = FixSizeOrderedDict(max=row.memory_size)
a.tumor_movement_stopping_rate = row.tumor_movement_stopping_rate
self.agents.append(a)
def make_nanoagents_from_scratch(self, cure_agent_type, radoznalost,
agent_memory_type, agent_memory_range,
tumor_movement_stopping_range, grid_size):
from itertools import cycle
self.xprint("Making nanoagents from scratch")
positions = cycle(self.get_random_positions_on_empty_cells())
for i in range(self.cure_number):
pos = next(positions)
self.xprint(pos)
a = cure_agent_type(
uuid.uuid4(),
self,
speeds=self.speeds,
radoznalost=radoznalost,
memory_type=agent_memory_type,
memory_range=agent_memory_range,
tumor_movement_stopping_range=tumor_movement_stopping_range,
probabilities_ranges=self.probabilites_ranges)
self.grid.place_agent(a, pos)
self.schedule.add(a)
def simulation_mode_function(self):
self.xprint("In simulation mode - not duplicating or mutating")
return
def generate_cancer_cell_positions(self, grid_size, cancer_cells_number):
center = grid_size // 2
poss = [(center, center)]
for pos in poss:
poss += [
n for n in self.grid.get_neighborhood(
pos, moore=True, include_center=False) if n not in poss
]
if len(set(poss)) >= cancer_cells_number:
break
poss = list(set(poss))
return poss
def duplicate_mutate_or_kill(self):
if self.mode == "simulation":
assert (False)
koliko = math.ceil(
percentage(self.MUTATION_PERCENTAGE, self.cure_number))
cureagents = [
c for c in self.schedule.agents if isinstance(c, CureAgent)
]
sortirani = sorted(cureagents, key=lambda x: x.points, reverse=True)
poslednji = sortirani[-koliko:]
prvi = sortirani[:koliko]
sredina = len(sortirani) // 2
pocetak_sredine = sredina - (koliko // 2)
kraj_sredine = sredina + (koliko // 2)
srednji = sortirani[pocetak_sredine:kraj_sredine]
self.mutate_agents(srednji)
assert (len(prvi) == len(poslednji))
self.remove_agents(poslednji)
self.duplicate_agents(prvi)
def mutate_agents(self, agents):
self.xprint("Mutating middle agents")
for a in agents:
a.mutate()
def remove_agents(self, agents):
for a in agents:
self.kill_cell(a)
def duplicate_agents(self, agents):
for a in agents:
a_new = a.copy()
self.grid.place_agent(a_new, (1, 1))
self.schedule.add(a_new)
def kill_cell(self, cell):
self.grid.remove_agent(cell)
self.schedule.remove(cell)
def detach_stem_cell(self, cell):
self.metastasis_score += 1
self.kill_cell(cell)
def kill_all_agents(self):
agents = [a for a in self.schedule.agents if isinstance(a, CureAgent)]
for a in agents:
a.kill_self()
def write_population_to_file(self, i):
df = record_model_population(self)
df.to_csv("./Populations/Population{}-step{}.csv".format(
self.SAMPLE_i, i))
def step(self):
self.datacollector.collect(self)
if self.mode == "simulation":
self.simulation_step()
self.write_population_to_file(self.counter)
self.schedule.step()
self.counter += 1
self.lifetime_counter += 1
if self.counter % self.STEPS_BEFORE_MUTATION == 0:
#ovde ga stavljamo da izbegnemo da na nultom koraku uradi to
self.duplicate_mutate_or_kill()
def simulation_step(self):
LIFETIME_OF_NANOAGENTS = 80
REINJECTION_PERIOD = 100
if self.lifetime_counter > LIFETIME_OF_NANOAGENTS:
self.kill_all_agents()
# if self.counter%(LIFETIME_OF_NANOAGENTS+REINJECTION_PERIOD)==0:
# #svakih 180 koraka se ubrizgavaju
# self.inject_nanoagents()
# self.lifetime_counter = 0
agents_at_each_step = round(self.cure_number / 14)
for i in range(agents_at_each_step):
try:
self.schedule.add(next(self.agents_iterator))
except StopIteration:
break
agents = [a for a in self.schedule.agents if isinstance(a, CureAgent)]
assert (len(agents) <= self.cure_number)
class World(Model):
def __init__(self, N, coop, e_prob, width=100, height=100):
self.num_agents = N
self.grid = MultiGrid(width, height, False)
self.schedule = RandomActivation(self)
self.running = True
self.population_center_x = np.random.randint(POP_MARGIN, self.grid.width - POP_MARGIN)
self.population_center_y = np.random.randint(POP_MARGIN, self.grid.height - POP_MARGIN)
self.num_energy_resources = RESERVE_SIZE
self.num_explorers = 0
self.num_exploiters = 0
self.datacollector = DataCollector(model_reporters={"Num_Explorer": "num_explorers","Num_Exploiter": "num_exploiters"})
self.bases = self.init_base()
self.ages = dict()
self.memoryLens = dict() # This will store average memory length at death for agent
self.expected_ages = []
self.member_tracker = []
self.energy_tracker = []
self.decay_rates = self.init_decay_rates()
# add social agents to the world
for i in range(1, self.num_agents + 1):
if np.random.uniform() <= EXPLORER_RATIO:
# create a new explorer
a = Explorer("explorer_%d" %(i), self, coop)
self.num_explorers += 1
else:
# create a new exploiter
a = Exploiter("exploiter_%d" %(i), self, coop, e_prob)
self.num_exploiters += 1
# keep society members confined at beginning
x = np.random.randint(self.population_center_x - POP_SPREAD, self.population_center_x + POP_SPREAD)
y = np.random.randint(self.population_center_y - POP_SPREAD, self.population_center_y + POP_SPREAD)
self.grid.place_agent(a, (x, y))
# add agent to scheduler
self.schedule.add(a)
self.expected_ages.append(a.energy / a.living_cost)
# add energy reserves to the world
for i in range(self.num_energy_resources):
a = EnergyResource("energy_reserve_%d" %(i), self, self.decay_rates[i])
# decide location of energy reserve
x = np.random.randint(0, self.grid.width)
y = np.random.randint(0, self.grid.height)
self.grid.place_agent(a, (x, y))
self.schedule.add(a)
def init_base(self):
pos = (self.population_center_x, self.population_center_y)
return self.grid.get_neighborhood(pos, True, radius=POP_SPREAD)
def init_decay_rates(self):
decay_rates = list(np.random.uniform(-0.1, 0.02, self.num_energy_resources))
np.random.shuffle(decay_rates)
return decay_rates
def step(self):
self.datacollector.collect(self)
self.schedule.step()
self.member_tracker.append((self.num_explorers, self.num_exploiters))
# keep track of total energy in world
if self.schedule.time % 50 == 0:
energies = [e.reserve for e in self.schedule.agents if isinstance(e, EnergyResource)]
if len(energies) > 0:
mean_energy = np.mean(energies)
else:
mean_energy = 0
self.energy_tracker.append(mean_energy)
# change location of energy resources every 500 steps randomly
# and change decay_rate to opposite every 100 steps
if self.schedule.time % 50 == 0:
for e in self.schedule.agents:
if isinstance(e, EnergyResource) and np.random.uniform() < 0.1:
# change location
e.decay_rate *= -1
if self.schedule.time % 500 == 0:
self.grid.remove_agent(e)
x = np.random.randint(0, self.grid.width)
y = np.random.randint(0, self.grid.height)
self.grid.place_agent(e, (x, y))
class World(Model):
def __init__(self, N, coop, e_prob, width=100, height=100):
self.num_agents = N
self.grid = MultiGrid(width, height, False)
self.schedule = RandomActivation(self)
self.running = True
self.population_center_x = np.random.randint(
POP_MARGIN, self.grid.width - POP_MARGIN)
self.population_center_y = np.random.randint(
POP_MARGIN, self.grid.height - POP_MARGIN)
self.num_energy_resources = RESERVE_SIZE
self.num_explorers = 0
self.num_exploiters = 0
self.datacollector = DataCollector(model_reporters={
"Num_Explorer": "num_explorers",
"Num_Exploiter": "num_exploiters"
})
self.bases = self.init_base()
self.ages = []
self.expected_ages = []
self.member_tracker = []
self.energy_tracker = []
# add social agents to the world
for i in range(1, self.num_agents + 1):
if np.random.uniform() <= EXPLORER_RATIO:
# create a new explorer
a = Explorer("explorer_%d" % (i), self, coop)
self.num_explorers += 1
else:
# create a new exploiter
a = Exploiter("exploiter_%d" % (i), self, coop, e_prob)
self.num_exploiters += 1
# keep society members confined at beginning
x = np.random.randint(self.population_center_x - POP_SPREAD,
self.population_center_x + POP_SPREAD)
y = np.random.randint(self.population_center_y - POP_SPREAD,
self.population_center_y + POP_SPREAD)
self.grid.place_agent(a, (x, y))
# add agent to scheduler
self.schedule.add(a)
self.expected_ages.append(a.energy / a.living_cost)
# add energy reserves to the world
for i in range(self.num_energy_resources):
a = EnergyResource("energy_reserve_%d" % (i), self)
# decide location of energy reserve
x = np.random.randint(0, self.grid.width)
y = np.random.randint(0, self.grid.height)
self.grid.place_agent(a, (x, y))
self.schedule.add(a)
def init_base(self):
pos = (self.population_center_x, self.population_center_y)
return self.grid.get_neighborhood(pos, True, radius=POP_SPREAD)
def step(self):
self.datacollector.collect(self)
self.schedule.step()
self.member_tracker.append((self.num_explorers, self.num_exploiters))
# keep track of total energy in world
energies = [
e.reserve for e in self.schedule.agents
if isinstance(e, EnergyResource)
]
if len(energies) > 0:
mean_energy = np.mean(energies)
else:
mean_energy = 0
self.energy_tracker.append(mean_energy)
# adjust decay rates to keep mean energy balanced in the world
if mean_energy <= MEAN_RESERVE - 5 * STDDEV_ENERGY:
for e in self.schedule.agents:
if isinstance(e, EnergyResource) and np.random.uniform() < 0.1:
e.decay_rate -= DECAY_RATE_ADJUST
elif mean_energy >= MEAN_RESERVE + 5 * STDDEV_RESERVE:
for e in self.schedule.agents:
if isinstance(e, EnergyResource) and np.random.uniform() < 0.1:
e.decay_rate += DECAY_RATE_ADJUST
# change location of energy resources every 100 steps randomly
if self.schedule.time % 100 == 0:
for e in self.schedule.agents:
if isinstance(e, EnergyResource) and np.random.uniform() < 0.1:
# change location
self.grid.remove_agent(e)
x = np.random.randint(0, self.grid.width)
y = np.random.randint(0, self.grid.height)
self.grid.place_agent(e, (x, y))
class PsyRTSGame(Model):
# id generator to track run number in batch run data
id_gen = itertools.count(1)
height = 20
width = 20
verbose = False # Print-monitoring
description = 'A model for simulating participants running the different experiments.'
def __init__(self, height=20, width=20, visibility = False ,initial_competitors=0,
initial_explorers=1, initial_predators=0 ,
impactTotalVisibility =.3 , impactPartialVisibility = .35, impactParticipants = .0,
impactCompetitors= .0, impactPredators= .0 ):
'''
Create a new PsyRTS model with the given parameters.
Args
initial_competitors: Number of units each participant knows
initial_explorers: Number of sheep to start with
initial_predators: Number of wolves to start with
visibility : total or partial visibility
'''
super().__init__()
self.uid = next(self.id_gen)
self.fps = 0
# Set parameters
self.height = height
self.width = width
self.visibility = visibility
self.initial_competitors = initial_competitors
self.initial_explorers = initial_explorers
self.initial_predators = initial_predators
self.schedule = RandomActivationByBreed(self)
self.grid = MultiGrid(self.height, self.width, torus=False)
self.uncertainty = 0.0
self.resources= 0
self.resourcesParticipants =0
self.resourcesCompetitors = 0
self.visitedCells = 0
self.stepsExploring = 0
self.stepsExploiting = 0
#parameters model
mu, sigma = impactTotalVisibility, 0.05 # mean and standard deviation
tv = np.random.normal(mu, sigma)
if tv < 0:
tv = 0
# print( "Starting with tv " , tv)
mu, sigma = impactPartialVisibility, 0.05 # mean and standard deviation
pv = np.random.normal(mu, sigma)
if pv < 0:
pv = 0
#print("Starting with pv ", pv)
self.impactTotalVisibility = tv
self.impactPartialVisibility = pv
self.impactParticipants = impactParticipants
self.impactCompetitors = impactCompetitors
self.impactPredators = impactPredators
self.TotalCells = next_moves = self.grid.get_neighborhood( (10,10), True, True, 11)
locationsResources = [(4,3) , (16,3), (3,10) , (10,10),(17,10) , (16,17),(4,17) ]
locationCPCompetitor = (10, 3)
locationCPParticipant = (10, 17)
self.locationsExploitation =[]
for loc in locationsResources:#resources
self.locationsExploitation = self.locationsExploitation + self.grid.get_neighborhood(loc, True, True,1)
#cfp
self.locationsExploitation = self.locationsExploitation + self.grid.get_neighborhood(locationCPParticipant, True, True, 1)
#print( "size grid exploiit " + str(len(self.locationsExploitation)))
self.locationsExploration = list(set(self.TotalCells) - set(self.locationsExploitation))
#print("size grid explore " + str(len(self.locationsExploration)))
locationsParticipants = [(10, 16), (13, 15), (7, 15), ( 8, 15), (12, 15) ]
locationsCompetitors = [(10, 4), (13, 5), (7, 5), ( 8, 5), (12, 5) ]
locationsPredators = [(10, 10), (13, 10), (7, 10), ( 8, 10), (12, 10) ]
self.datacollector = DataCollector(
model_reporters={
"Experiment_Synth": track_experiment,
"Conditions": track_params,
"Step": track_run,
"Exploration": exploration ,
"Exploitation": exploitation,
"ResourcesRatio": resourcesRatio,
"ProportionEE": proportionEE,
"MapExplored": mapExplored,
}) #reporto a datos
centralplaceparticipant = CentralPlace(self.next_id(), locationCPParticipant, self, True)
self.grid.place_agent(centralplaceparticipant, locationCPParticipant)
self.schedule.add(centralplaceparticipant)
centralplacecompetitor = CentralPlace(self.next_id(), locationCPCompetitor, self, False)
self.grid.place_agent(centralplacecompetitor, locationCPCompetitor)
self.schedule.add(centralplacecompetitor)
# Create competitor:
for i in range(self.initial_competitors):
sheep = Competitor(self.next_id(), locationsCompetitors[i], self, True, centralplacecompetitor)
self.grid.place_agent(sheep, locationsCompetitors[i])
self.schedule.add(sheep)
# Create explorers:
for i in range(self.initial_explorers):
sheep = Participant(self.next_id(), locationsParticipants[i], self, True, centralplaceparticipant)
self.grid.place_agent(sheep, locationsParticipants[i])
self.schedule.add(sheep)
# Create predators
for i in range(self.initial_predators):
wolf = Predator(self.next_id(), locationsPredators[i], self, True)
self.grid.place_agent(wolf, locationsPredators[i])
self.schedule.add(wolf)
for pa in locationsResources:
# randrange gives you an integral value
irand = randrange(1, 11)
#irand = 4
#print("resource con valores {} ". format( irand) )
self.resources = self.resources + irand
patch = Resources(self.next_id(), pa, self, irand)
self.grid.place_agent(patch, pa)
self.schedule.add(patch)
for x in range(0,20):
for y in range(0, 20):
breadcrumb = BreadCrumb(self.next_id(), (x,y), self)
self.grid.place_agent(breadcrumb, (x,y))
self.schedule.add(breadcrumb)
self.running = True
self.datacollector.collect(self)
def updateUncertainty(self):
# information que gana por conocer el ambiente
#participantExploration = (number_visited(self, True) +number_visited(self, False))/400
participantExploration = mapExplored(self)
uncertaintyVisibility = 0
# if self.model.visibility :
# multitaskingFriction = (1 - self.model.impactParticipants ) ** self.model.initial_explorers
# else:
if self.visibility:
uncertaintyVisibility = self.impactTotalVisibility
self.uncertainty = uncertaintyVisibility
else:
uncertaintyVisibility = self.impactPartialVisibility
self.uncertainty = uncertaintyVisibility* (1-participantExploration)
if self.uncertainty <0 :
self.uncertainty = 0
#print("explored map {} uncertainty in the environment {} ".format ( participantExploration, self.uncertainty ))
def step(self):
self.datacollector.collect(self)
self.schedule.step()
self.updateUncertainty()
# collect data
participantsAlive = self.schedule.get_breed_count(Participant)
if participantsAlive <=0:
print("Stop Participants dead")
self.running = False
if self.resourcesCompetitors +self.resourcesParticipants == self.resources:
print("Stop No More Resources")
self.running = False
if self.schedule.steps>150:
print("toomany steps")
self.running = False
def run_model(self, step_count=150):
for i in range(step_count):
self.step()