class Bootstrap(object):
def __init__(self, env):
self.env = env
# create simulation enviromment
img = np.zeros((900, 800, 3), dtype=np.uint8)
self.sim = Simulation(self.env, img)
def processSimulation(self):
# initialize Sioux Falls network
siouxfalls = SiouxFalls(0.0025)
# create Sioux Fall network by enumerating across all links
for linkid, t0 in enumerate(siouxfalls.t0):
# calculate length of link with sqrt((x1 - x2)^2 + (y1 - y2)^2)
length = np.sqrt(
np.power(siouxfalls.x1[linkid] - siouxfalls.x2[linkid], 2)
+ np.power(siouxfalls.y1[linkid] - siouxfalls.y2[linkid], 2)) / 600.
mu = siouxfalls.mu[linkid]
# assign nodeID to each link if check pass in node list
for i, node in enumerate(siouxfalls.nodes):
if linkid+1 in node:
nodeID = i
# assign turn ratio to each link
turns = {}
for j, turn in enumerate(siouxfalls.turns[linkid]):
turns[j + 1] = turn
# assign exit probability from last item in turn list ([-1])
turns['exit'] = turns.pop(list(turns.keys())[-1])
# generate coordinates of each link (for visualization)
pt1 = (np.float32(siouxfalls.x1[linkid] / 600.),
np.float32(siouxfalls.y1[linkid] / 600.))
pt2 = (np.float32(siouxfalls.x2[linkid] / 600.),
np.float32(siouxfalls.y2[linkid] / 600.))
c = (pt1, pt2)
# draw link on map
self.sim.networkLines.append(c)
# add link to sim.network
self.sim.network.addLink(linkID=linkid+1, turns=turns,
type='link', length=length,
t0=t0, MU=mu, nodeID=nodeID,
coordinates=c)
# initialize car generation
self.env.process(self.sim.source(
10, LAMBDA=siouxfalls.flambda[linkid], linkid=linkid+1))
yield self.env.timeout(1)
def main():
#################################
# initialize discrete event env #
#################################
env = simpy.Environment() # use instant simulation
# env = simpy.rt.RealtimeEnvironment(factor=1.) # use real time simulation
# initialize Sioux Falls network
siouxfalls = SiouxFalls(0.0025)
# create simulation enviromment
img = np.zeros((900, 800, 3), dtype=np.uint8)
sim = Simulation(env, img)
# create Sioux Fall network by enumerating across all links
for linkid, t0 in enumerate(siouxfalls.t0):
# calculate length of link with sqrt((x1 - x2)^2 + (y1 - y2)^2)
length = np.sqrt(
np.power(siouxfalls.x1[linkid] - siouxfalls.x2[linkid], 2) +
np.power(siouxfalls.y1[linkid] - siouxfalls.y2[linkid], 2)) / 600.
mu = siouxfalls.mu[linkid]
# assign nodeID to each link if check pass in node list
for i, node in enumerate(siouxfalls.nodes):
if linkid + 1 in node:
nodeID = i
# assign turn ratio to each link
turns = {}
for j, turn in enumerate(siouxfalls.turns[linkid]):
turns[j + 1] = turn
# assign exit probability from last item in turn list ([-1])
turns['exit'] = turns.pop(list(turns.keys())[-1])
# generate coordinates of each link (for visualization)
pt1 = (np.float32(siouxfalls.x1[linkid] / 600.),
np.float32(siouxfalls.y1[linkid] / 600.))
pt2 = (np.float32(siouxfalls.x2[linkid] / 600.),
np.float32(siouxfalls.y2[linkid] / 600.))
c = (pt1, pt2)
# draw link on map
sim.networkLines.append(c)
# add link to sim.network
sim.network.addLink(linkID=linkid + 1,
turns=turns,
type='link',
length=length,
t0=t0,
MU=mu,
nodeID=nodeID,
coordinates=c)
# initialize car generation
env.process(
sim.source(10,
LAMBDA=siouxfalls.flambda[linkid],
linkid=linkid + 1))
# draw initial network
for i in sim.networkLines:
cv2.line(sim.img, i[0], i[1], (255, 255, 255), 3)
for linkid in sim.network.links:
loc = (0.25 * np.asarray(sim.network.links[linkid]['coordinates'][1]) +
0.75 * np.asarray(sim.network.links[linkid]['coordinates'][0]))
cv2.putText(sim.img, str(linkid), tuple(loc), cv2.FONT_HERSHEY_SIMPLEX,
0.5, (255, 255, 255))
name = 'Sioux Falls Network'
cv2.imshow(name, sim.img)
# start visualization update process
env.process(sim.visualization(frequency=0.2, name=name))
# wait for keypress to start simulation
print('press space to start')
k = cv2.waitKey(0)
if k == 27:
sys.exit()
# run simulation
env.run()
######################################
# simulation statistics and graphing #
######################################
df = pd.DataFrame(
sorted(sim.data, key=lambda x: x[3]),
columns=['carID', 'link', 'event', 'time', 'queue', 't_queue'])
print(df)
# cars statistics
totalTravelTime = (df[['carID', 'time']].groupby(['carID']).max() -
df[['carID', 'time']].groupby(['carID']).min())
totalTravelTime.columns = ['totalTravelTime']
totalSegments = df.loc[df['event'] == 'arrival'][['carID', 'link']]
totalSegments = totalSegments.groupby(['carID']).count()
totalSegments.columns = ['totalSegments']
meanWaitTime = df.loc[df['event'] == 'departure'][['carID', 't_queue']]
meanWaitTime = meanWaitTime.groupby(['carID']).mean()
meanWaitTime.columns = ['meanWaitTime']
carStatistics = pd.concat([totalTravelTime, totalSegments, meanWaitTime],
axis=1)
# links statistics
stats.meanQueueLength(plt, df)
plt.figure(2)
for link in sim.network.links.keys():
df2 = df.loc[(df['link'] == link) & (df['event'] != 'entry')]
if df2.empty is False and df2['t_queue'].sum() > 0.:
plt.plot(df2[['time']], df2[['queue']], label='link %s' % link)
plt.title('Queueing Simulation')
plt.ylabel('queue length')
plt.xlabel('time (s)')
plt.legend()
plt.show()
print('Press any key to exit')
cv2.waitKey(0)
