def table(self):
configuration = self.graph.getConfiguration()
sanitise = lambda x:str(util.ci(x))
# Include those variables which vary in the graph.
headingkeys = [ x for x in config.configvariables.keys() \
if x not in configuration.keys() ]
# Create a row for each data point.
rows = []
for x in self.graph.data:
heading = [ "%s %s" % (util.ci(getattr(x, k)), config.configvariables[k]) \
for k in headingkeys ]
heading = string.join(heading)
colour = (x.better() and "green") or \
(x.significant() and "red") or "white"
row = TABLEROW % (colour,
heading,
x.productdata.mean,
x.productdata.count,
x.referencedata.mean,
x.referencedata.count,
x.relative)
row = re.sub("_", "\_", row)
rows.append(row)
rows = string.join(rows, "\\\\ \hline\n")
rows += "\\\\ \hline\n"
return TABLE % rows
def average_froc_with_ci(frocs, fppi_range, confidence=0.95):
av_sen = []
for i in range(len(frocs)):
x = frocs[i].T
f = interp1d(x[0],
x[1],
kind='linear',
fill_value=0.0,
bounds_error=False)
sen = f(fppi_range)
av_sen.append(sen)
av_sen = np.array(av_sen)
assert len(fppi_range) == len(av_sen[i])
aucs = []
for i in range(len(av_sen)):
froc = np.array([fppi_range, av_sen[i]]).T
aucs.append(util.auc(froc, fppi_range))
lo, up = util.ci(aucs, confidence)
av_froc = [fppi_range, np.mean(av_sen, axis=0), np.std(av_sen, axis=0)]
av_froc = np.array(av_froc).T
return av_froc, lo, up
def create(self):
self.figure = matplotlib.pyplot.figure()
axis = self.figure.add_subplot(1, 1, 1)
errorvalues = [x.error for x in self.data]
productvalues = [x.productdata.mean for x in self.data]
referencevalues = [x.referencedata.mean for x in self.data]
index = numpy.arange(len(productvalues))
# Work out the colours to use for the product bars.
colours = []
for x in self.data:
if x.better():
colours.append(self.bettercolour)
elif x.significant():
colours.append(self.worsecolour)
else:
colours.append(self.colour)
# Plot the product bars.
axis.bar(
index,
productvalues,
self.width,
yerr=errorvalues,
color=colours,
ecolor=self.errorcolour,
linewidth=self.linewidth,
)
# Plot the reference bars next to the product bars.
axis.bar(
index + self.width,
referencevalues,
self.width,
color=self.referencecolour,
ecolor=self.errorcolour,
linewidth=self.linewidth,
)
# Set the axes' ranges.
axis.set_ylim(ymin=0)
# Spruce the graph up a bit.
xlabels = [
string.join(["%s %s" % (util.ci(getattr(x, y)), config.configvariables[y]) for y in config.bars], "\n")
for x in self.data
]
axis.set_xticklabels(xlabels, size=self.labelpointsize)
axis.set_xticks(index + self.width)
axis.set_ylabel(self.data[0].units)
for x in axis.get_ymajorticklabels():
x.set_size(self.labelpointsize)
axis.set_clip_on(False)
def create(self):
self.figure = matplotlib.pyplot.figure()
axis = self.figure.add_subplot(1, 1, 1)
errorvalues = [x.error for x in self.data]
productvalues = [x.productdata.mean for x in self.data]
referencevalues = [x.referencedata.mean for x in self.data]
index = numpy.arange(len(productvalues))
# Work out the colours to use for the product bars.
colours = []
for x in self.data:
if x.better(): colours.append(self.bettercolour)
elif x.significant(): colours.append(self.worsecolour)
else: colours.append(self.colour)
# Plot the product bars.
axis.bar(index,
productvalues,
self.width,
yerr=errorvalues,
color=colours,
ecolor=self.errorcolour,
linewidth=self.linewidth)
# Plot the reference bars next to the product bars.
axis.bar(index + self.width,
referencevalues,
self.width,
color=self.referencecolour,
ecolor=self.errorcolour,
linewidth=self.linewidth)
# Set the axes' ranges.
axis.set_ylim(ymin=0)
# Spruce the graph up a bit.
xlabels = [ string.join([ "%s %s" % (util.ci(getattr(x, y)), config.configvariables[y]) \
for y in config.bars ], "\n") \
for x in self.data ]
axis.set_xticklabels(xlabels, size=self.labelpointsize)
axis.set_xticks(index + self.width)
axis.set_ylabel(self.data[0].units)
for x in axis.get_ymajorticklabels():
x.set_size(self.labelpointsize)
axis.set_clip_on(False)
def include(self, level=""):
level = "sub" * len(config.documentsections)
configuration = self.graph.getConfiguration()
# Function to clean up the config strings.
sanitise = lambda x:str(util.ci(x))
caption = []
for k in config.displayorder:
if not configuration.has_key(k):
continue
configuration[k] = sanitise(configuration[k])
# Don't repeat information in the report.
#if not k in config.documentsections:
caption.append("%s %s" % (string.upper(str(configuration[k])),
config.configvariables[k]))
caption = string.join(caption)
caption = re.sub("_", "-", caption)
return INCLUDE % (level,
caption,
self.graph.filename,
self.path,
self.graph.filename,
caption,
self.path,
self.graph.filename)
def perf_measure_1():
# 1. For individual vehicles: distribution of the travel time (mean, standard deviation, plot
# a histogram). Indicate which part of the travel time is caused by the delay. Present
# results on the number of incidents they encounter on average (again: mean, standard
# deviation, distribution).
# Travel time on the route between A to B
# Probability of minutes of time taken for the travel along the route. Ghost cars not necessary, or included.
graph_type = 'Highway'
system = run_simulation('Measure-1',
seed=30,
end_time=1600,
stats_step_size=50,
graph_type=graph_type,
tow_truck_mode=False,
warm_up=100)
travel_times_list = system.final_statistics['route_travel_times']
delay_times_list = system.final_statistics['route_delay_times']
number_delays_list = system.final_statistics['route_number_delays']
print('Travel Times List for Route Cars')
print(travel_times_list)
print(len(travel_times_list))
# Graph 1
rounded_travel_times_list = [round(i) for i in travel_times_list]
max_travel_time = max(rounded_travel_times_list)
counts_travel_times = Counter(rounded_travel_times_list)
list_of_times = []
print('Rounded Travel Times List', rounded_travel_times_list)
for i in range(int(max_travel_time) + 1):
list_of_times.append(counts_travel_times[i])
print(list_of_times) # Must be X-axis
print('Count of travel times -', list_of_times)
plot = plt.figure()
plt.ylabel('Probability')
plt.xlabel('Total Travel Time')
plt.xlim(50, max_travel_time)
plt.plot([i for i in range(len(list_of_times))],
[j / sum(list_of_times) for j in list_of_times])
plt.show()
# Graph 2
rounded_delay_times_list = [round(i) for i in delay_times_list]
max_delay_time = max(rounded_delay_times_list)
counts_delay_times = Counter(rounded_delay_times_list)
list_of_delay_times = []
print('Rounded Delay Times List', rounded_delay_times_list)
for i in range(1, int(max_delay_time) + 1):
list_of_delay_times.append(counts_delay_times[i])
print(list_of_delay_times) # Must be X-axis
print('Count of delay times -', list_of_delay_times)
plot = plt.figure()
plt.ylabel('Probability')
plt.xlabel('Total Delay Time')
plt.xlim(1, max_delay_time)
plt.plot([i for i in range(len(list_of_delay_times))],
[j / sum(list_of_delay_times) for j in list_of_delay_times])
plt.show()
# Graph 3
# Showing the distribution of the number of incidents in a single route travel
print(number_delays_list)
counts_number_delays = Counter(number_delays_list)
list_of_number_delays = []
for i in range(max(number_delays_list) + 1):
list_of_number_delays.append(counts_number_delays[i])
print('Counts of number of delays', list_of_number_delays)
plot = plt.figure()
plt.xlabel('Numbers of incidents')
plt.ylabel('Probability')
# plt.xlim(1, max_delay_time)
plt.bar([i for i in range(len(list_of_number_delays))],
[j / sum(list_of_number_delays) for j in list_of_number_delays])
plt.show()
# CI calculation
travel_times_list = []
delay_times_list = []
number_delays_list = []
for i in range(run_size):
system = run_simulation('Measure-1',
seed=None,
end_time=1600,
stats_step_size=50,
graph_type=graph_type,
tow_truck_mode=False,
warm_up=100)
travel_times_list += system.final_statistics['route_travel_times']
delay_times_list += system.final_statistics['route_delay_times']
number_delays_list += system.final_statistics['route_number_delays']
print('\n')
print('Number of Route Travel Times - ', len(travel_times_list))
mean_route_travel_time = sum(travel_times_list) / len(travel_times_list)
std_route_travel_time = std(travel_times_list)
print('Mean of Route Travel Time -', mean_route_travel_time)
print('Standard Deviation of Route Travel Time -', std_route_travel_time)
halfwidth_route_travel_time = ci(std_route_travel_time, run_size)
print('Halfwidth of Route Travel Time -', halfwidth_route_travel_time)
print('Range - ', mean_route_travel_time - halfwidth_route_travel_time,
mean_route_travel_time + halfwidth_route_travel_time)
# Also do CI calculation
print('\n')
print('Number of Delay Travel Times - ', len(delay_times_list))
mean_delay_time = sum(delay_times_list) / len(delay_times_list)
std_delay_time = std(delay_times_list)
print('Mean of Delay Travel Time -', mean_delay_time)
print('Standard Deviation of Delay Travel Time -', std_delay_time)
halfwidth_delay_time = ci(std_delay_time, run_size)
print('Halfwidth of Delay Travel Time -', halfwidth_delay_time)
print('Range - ', mean_delay_time - halfwidth_delay_time,
mean_delay_time + halfwidth_delay_time)
print('\n')
print('Number of Delays - ', len(number_delays_list))
mean_delay_number = sum(number_delays_list) / len(number_delays_list)
std_delay_number = std(number_delays_list)
print('Mean of Delay Travel Time -', mean_delay_number)
print('Standard Deviation of Delay Travel Time -', std_delay_number)
halfwidth_delay_number = ci(std_delay_number, run_size)
print('Halfwidth of Delay Travel Time -', halfwidth_delay_number)
print('Range - ', mean_delay_number - halfwidth_delay_number,
mean_delay_number + halfwidth_delay_number)
def perf_measure_3():
# For incidents: distribution of the number of incidents in the whole network, at any
# arbitrary epoch. This is similar to the previous question, but now for the fraction of time
# that exactly k incidents are taking place simultaneously in the network
graph_type = 'Highway'
system = run_simulation('Measure-3-vis',
seed=40,
end_time=1600,
stats_step_size=50,
graph_type=graph_type,
tow_truck_mode=False,
warm_up=100)
times_fraction_per_number_of_incidents = system.final_statistics[
'time_spent_per_current_number_of_incidents']
# rounded_times = [round(i) for i in times_fraction_per_number_of_incidents]
# print(rounded_times)
# counts_number_delays = Counter(number_delays_list)
# list_of_number_delays = []
# for i in range(max(number_delays_list) + 1):
# list_of_number_delays.append(counts_number_delays[i])
print('Counts of time spent per number of incidents',
times_fraction_per_number_of_incidents)
print('len', len(times_fraction_per_number_of_incidents))
min_non_zero_incidents = 0
max_non_zero_incidents = len(times_fraction_per_number_of_incidents)
i = 0
while i < len(times_fraction_per_number_of_incidents
) and times_fraction_per_number_of_incidents[i] == 0:
min_non_zero_incidents += 1
i += 1
i = len(times_fraction_per_number_of_incidents) - 1
while i > 0 and times_fraction_per_number_of_incidents[i] == 0:
max_non_zero_incidents -= 1
i -= 1
print('max_non_zero_incidents', max_non_zero_incidents)
print('min_non_zero_incidents', min_non_zero_incidents)
plot = plt.figure()
plt.xlabel('Numbers of incidents')
plt.ylabel('Fraction of time spent')
plt.xlim(min_non_zero_incidents - 1, max_non_zero_incidents)
plt.bar([i for i in range(len(times_fraction_per_number_of_incidents))], [
j / sum(times_fraction_per_number_of_incidents)
for j in times_fraction_per_number_of_incidents
])
plt.show()
average_number_of_incidents = []
# CI calculation
for i in range(run_size):
system = run_simulation('Measure-3' + str(i),
seed=i,
end_time=1600,
stats_step_size=50,
graph_type=graph_type,
tow_truck_mode=False,
warm_up=100)
print('Run -', i)
average_number_of_incidents.append(
system.final_statistics['average_number_of_incidents'])
mean_number_of_incidents = sum(average_number_of_incidents) / len(
average_number_of_incidents)
std_number_of_incidents = std(average_number_of_incidents)
print('Mean of Number of Incidents -', mean_number_of_incidents)
print('Standard Deviation of Number of Incidents -',
std_number_of_incidents)
halfwidth_number_of_incidents = ci(std_number_of_incidents, run_size)
print('Halfwidth of Number of Incidents -', halfwidth_number_of_incidents)
print('Range - ', mean_number_of_incidents - halfwidth_number_of_incidents,
mean_number_of_incidents + halfwidth_number_of_incidents)
def perf_measure_2():
# 2. For the network: distribution of the number of delayed vehicles at any arbitrary epoch.
# Formulated differently, we would like to see a table where you simulate for k = 0; 1; 2; : : :
# the fraction of time that exactly k vehicles are being delayed (consider it as being stuck
# in a traffic jam).
# Note that, in order to determine the second performance measure (number of
# delayed vehicles in the network), it actually matters how many vehicles are driving in the
# whole network. So now you need to add random vehicles in the network that drive between
# the highway junctions. This part does not have to be implemented very realistically. You
# may even assume that vehicles enter the network randomly at each of the nodes, travel one
# link, and then leave the network again. For these vehicles, no origin-destination routing is
# required. Only vehicles travelling from A to B have a fixed route. Our advice is to keep
# it simple: do not introduce too many model parameters, like number of vehicles per hour
# travelling on each link. You are allowed to assume that this rate is equal for all links.
graph_type = 'Highway'
system = run_simulation('Measure-2-vis',
seed=40,
end_time=1600,
stats_step_size=50,
graph_type=graph_type,
tow_truck_mode=False,
warm_up=100)
times_delayed_per_number_of_cars = system.final_statistics[
'time_spent_delayed_per_current_number_of_all_cars']
# print('Time spent per number of delayed cars', times_delayed_per_number_of_cars)
print('len', len(times_delayed_per_number_of_cars))
data = [(i, times_delayed_per_number_of_cars[i])
for i in range(len(times_delayed_per_number_of_cars))]
x, y, bin_width = group_into_bins(data)
print('x', x)
print('y', y)
print('len(y)', len(y))
min_non_zero_delayed_cars = 0
max_non_zero_delayed_cars = len(y) - 1
# comparison = 0
i = 0
while i < len(y) and y[i] == 0:
min_non_zero_delayed_cars += 1
i += 1
# i = x[i+1]
i = len(y) - 1
while i > 0 and y[i] == 0:
max_non_zero_delayed_cars -= 1
i -= 1
print('max_non_zero_delayed_cars', max_non_zero_delayed_cars)
print('min_non_zero_delayed_cars', min_non_zero_delayed_cars)
plt.xlabel('Numbers of delayed cars (with Tow Trucks)')
plt.ylabel('Fraction of time spent')
plt.xlim(x[min_non_zero_delayed_cars] - 1, x[max_non_zero_delayed_cars])
plt.bar(x, [j / sum(y) for j in y], width=bin_width * 0.8)
plt.show()
average_number_of_delayed_cars = []
# CI calculation
for i in range(run_size):
system = run_simulation('Measure-2' + str(i),
seed=i,
end_time=1600,
stats_step_size=50,
graph_type=graph_type,
tow_truck_mode=False,
warm_up=100)
print('Run -', i)
average_number_of_delayed_cars.append(
system.final_statistics['average_time_of_delay_of_all_cars'])
mean_number_of_delayed_cars = sum(average_number_of_delayed_cars) / len(
average_number_of_delayed_cars)
std_number_of_delayed_cars = std(average_number_of_delayed_cars)
print('Mean of Delayed Cars -', mean_number_of_delayed_cars)
print('Standard Deviation of Delayed Cars -', std_number_of_delayed_cars)
halfwidth_number_of_delayed_cars = ci(std_number_of_delayed_cars, run_size)
print('Halfwidth of Delayed Cars -', halfwidth_number_of_delayed_cars)
print('Range - ',
mean_number_of_delayed_cars - halfwidth_number_of_delayed_cars,
mean_number_of_delayed_cars + halfwidth_number_of_delayed_cars)