class Simulation:
def __init__(self,
num_individuals,
num_steps,
method="leap_frog",
tau=0.1,
v_des=1.5,
room="square",
room_size=25):
std_deviation = 0.07
variation = np.random.normal(
loc=1, scale=std_deviation, size=(1, num_individuals)
) # is late used to make the agents differ in weight and size
# Constants
self.L = room_size # size of square room (m)
self.N = num_individuals # quantity of pedestrians
self.tau = tau # time-step (s)
self.num_steps = num_steps # number of steps for integration
# Agent information
self.radii = 0.4 * (np.ones(self.N) *
variation).squeeze() # radii of pedestrians (m)
self.v_des = v_des * np.ones(self.N) # desired velocity (m/s)
self.m = 80 * (np.ones(self.N) *
variation).squeeze() # mass of pedestrians (kg)
self.forces = None # forces on the agents
self.agents_escaped = None #number of agents escaped by timesteps
self.v = np.zeros(
(2, self.N, self.num_steps)) # Three dimensional array of velocity
self.y = np.zeros(
(2, self.N, self.num_steps)
) # Three dimensional array of place: x = coordinates, y = Agent, z=Time
# other
self.room = Room(room,
room_size) # kind of room the simulation runs in
self.method = getattr(Integrators,
method) # method used for integration
self.diff_equ = Diff_Equ(self.N, self.L, self.tau, self.room,
self.radii,
self.m) # initialize Differential equation
# function set_time, set_steps give the possiblity to late change these variable when needed
def set_steps(self, steps):
self.num_steps = steps
# function to change the methode of integration if needed
def set_methode(self, method):
self.method = getattr(Integrators, method)
def dont_touch(
self, i, x
): # yields false if people don't touch each other and true if they do
for j in range(i - 1):
if np.linalg.norm(x - self.y[:, j, 0]) < 3 * self.radii[i]:
return True
return False
# fills the spawn zone with agents with random positions
def fill_room(self):
spawn = self.room.get_spawn_zone()
len_right = spawn[0, 1] - spawn[0, 0]
len_left = spawn[1, 1] - spawn[1, 0]
max_len = max(len_left, len_right)
# checks if the area is too small for the agents to fit in
area_people = 0
for i in range(self.N):
area_people += 4 * self.radii[i]**2
if area_people >= 0.7 * max_len**2:
sys.exit(
'Too much people! Please change the size of the room/spawn-zone or the amount of people.'
)
# checks if the agent touches another agent/wall and if so gives it a new random position in the spawn-zone
for i in range(self.N):
# The pedestrians don't touch the wall
x = len_right * np.random.rand() + spawn[0, 0]
y = len_left * np.random.rand() + spawn[1, 0]
pos = [x, y]
# The pedestrians don't touch each other
while self.dont_touch(i, x):
x = len_right * np.random.rand() + spawn[0, 0]
y = len_left * np.random.rand() + spawn[1, 0]
pos = [x, y]
self.y[:, i, 0] = pos
self.v[:, :, 0] = self.v_des * self.diff_equ.e_t(self.y[:, :, 0])
# calls the method of integration with the starting positions, diffequatial equation, number of steps, and delta t = tau
def run(self):
self.y, self.agents_escaped, self.forces = self.method(
self.y[:, :, 0], self.v[:, :, 0], self.diff_equ.f, self.num_steps,
self.tau, self.room)
# Displayes the simulation in pygame
def show(self, wait_time, sim_size):
display_graph(self.agents_escaped, self.forces, self.m)
display_events(self.y, self.room, wait_time, self.radii, sim_size,
self.agents_escaped)
