def main():
c = float(sys.argv[1])
multi = mx.read_multi(nodes_file_name = '2_multiplex/multiplex_unscaled_nodes.txt',
edges_file_name = '2_multiplex/multiplex_unscaled_edges.txt')
weights = ['uniform_time_m', 'free_flow_time_m']
print 'Means prior to rescaling'
for weight in weights:
x = np.array([multi.G.edge[e[0]][e[1]][weight] for e in multi.G.edges_iter()
if multi.G.edge[e[0]][e[1]]['layer'] == 'streets'])
print weight + ': ' + str(round(x.mean(),2))
print '\nMeans after rescaling'
for weight in weights:
multi.scale_edge_attribute(layer = 'streets', attribute = weight, beta = c)
x = np.array([multi.G.edge[e[0]][e[1]][weight] for e in multi.G.edges_iter()
if multi.G.edge[e[0]][e[1]]['layer'] == 'streets'])
print weight + ': ' + str(round(x.mean(),2))
multi.to_txt('2_multiplex/', 'mx')
def main():
no_metro_beta = 1000
betas = pd.read_csv('plot_betas.csv').beta
m = mx.read_multi(nodes_file_name = '3_throughput/mx_flow_nodes.txt',
edges_file_name = '3_throughput/mx_flow_edges.txt')
m.read_od(layer = 'taz', # keys are in this layer
key = 'taz', # this is the key attribute
od_file = '1_data/taz_od/0_1.txt', # here's where the file lives
sep = ' ') # this is what separates entries
mean_free_flow_time = m.mean_edge_attr_per(layers = ['streets'],
attr = 'free_flow_time_m',
weight_attr = 'flow_' + str(no_metro_beta))
mean_congested_time = m.mean_edge_attr_per(layers = ['streets'],
attr = 'congested_time_m_' + str(no_metro_beta),
weight_attr = 'flow_' + str(no_metro_beta))
dist = m.mean_edge_attr_per(layers = ['streets'],
attr = 'dist_km',
weight_attr = 'flow_' + str(no_metro_beta))
v_f = dist * 1.0 / mean_free_flow_time
v_c = dist * 1.0 / mean_congested_time
print 'Mean free flow speed = ' + str(round(v_f,2)) + 'km per minute'
print 'Mean congested speed = ' + str(round(v_c,2)) + 'km per minute'
# Scale metro so that it runs at v_c (this is beta = 1)
mean_metro_time = m.mean_edge_attr_per(layers = ['metro'],
attr = 'free_flow_time_m')
mean_metro_dist = m.mean_edge_attr_per(layers = ['metro'],
attr = 'dist_km')
v_m = mean_metro_dist / mean_metro_time
print 'Mean metro speed = ' + str(round(v_m,2)) + 'km per minute'
ratio = v_m / v_c
print 'Scaling metro speed by factor of ' + str(round(1 / ratio,2))
m.scale_edge_attribute(layer = 'metro',
attribute = 'free_flow_time_m',
beta = ratio)
print 'Metro now runs at mean speed of street layer'
# run
for beta in betas:
m.scale_edge_attribute(layer = 'metro',
attribute = 'free_flow_time_m',
beta = beta)
simulate(m, beta = beta, n = 50000)
m.scale_edge_attribute(layer = 'metro',
attribute = 'free_flow_time_m',
beta = 1.0/beta)
def main():
# Read in the multiplex
m = mx.read_multi()
# Read OD keyed to m
m.read_od(layer = 'taz',
key = 'taz',
od_file = '1_data/taz_od/0_1.txt',
sep = " ")
# compute ITA with no metro
no_metro_beta = 1000
ita_iteration(m, beta = no_metro_beta)
# compute the mean free flow speed v_f and the mean congested speed v_c
mean_free_flow_time = m.mean_edge_attr_per(layers = ['streets'],
attr = 'free_flow_time_m',
weight_attr = 'flow_' + str(no_metro_beta))
mean_congested_time = m.mean_edge_attr_per(layers = ['streets'],
attr = 'congested_time_m_' + str(no_metro_beta),
weight_attr = 'flow_' + str(no_metro_beta))
dist = m.mean_edge_attr_per(layers = ['streets'],
attr = 'dist_km',
weight_attr = 'flow_' + str(no_metro_beta))
v_f = dist * 1.0 / mean_free_flow_time
v_c = dist * 1.0 / mean_congested_time
print 'Mean free flow speed = ' + str(round(v_f,2)) + 'km per minute'
print 'Mean congested speed = ' + str(round(v_c,2)) + 'km per minute'
# Scale metro so that it runs at v_c (this is beta = 1)
mean_metro_time = m.mean_edge_attr_per(layers = ['metro'],
attr = 'free_flow_time_m')
mean_metro_dist = m.mean_edge_attr_per(layers = ['metro'],
attr = 'dist_km')
v_m = mean_metro_dist / mean_metro_time
print 'Mean metro speed = ' + str(round(v_m,2)) + 'km per minute'
ratio = v_m / v_c
print 'Scaling metro speed by factor of ' + str(round(1 / ratio,2))
m.scale_edge_attribute(layer = 'metro',
attribute = 'free_flow_time_m',
beta = ratio)
print 'Metro now runs at mean speed of street layer'
with open("true_beta.txt", "w") as text_file:
text_file.write('The true beta is approximately ' + str(1.0 / ratio))
# main loop: for each beta, scale the metro layer of m, run ITA, and save
# the results. Note that using summary = True prints out route-wise
# summaries for each level of beta, and is VERY computationally expensive.
betas = pd.read_csv('betas.csv').beta
for beta in betas:
ita_iteration(m, beta)
# start = time.clock()
# m.scale_edge_attribute(layer = 'metro',
# attribute = 'free_flow_time_m',
# beta = beta)
# df = m.run_ita(n_nodes = None,
# summary = True, # change this to get route tables
# attrname = 'congested_time_m_' + str(beta),
# flow_name = 'flow_' +str(beta),
# P = [.2, .2, .2, .2, .1, .1],
# scale = .25)
# if df is not None:
# df.to_csv('3_throughput/route_info_' + str(beta) + '.csv')
# m.scale_edge_attribute(layer = 'metro',
# attribute = 'free_flow_time_m',
# beta = 1.0/beta)
# time_taken = str(round((time.clock() - start) / 60.0, 1)) + 'm'
# print 'assignment for beta = ' + str(beta) + ' completed in ' + time_taken
m.to_txt('3_throughput/', 'mx_flow')
def main():
no_metro_beta = 1000
betas = pd.read_csv('plot_betas.csv').beta
m = mx.read_multi(nodes_file_name='3_throughput/mx_flow_nodes.txt',
edges_file_name='3_throughput/mx_flow_edges.txt')
m.read_od(
layer='taz', # keys are in this layer
key='taz', # this is the key attribute
od_file='1_data/taz_od/0_1.txt', # here's where the file lives
sep=' ') # this is what separates entries
mean_free_flow_time = m.mean_edge_attr_per(layers=['streets'],
attr='free_flow_time_m',
weight_attr='flow_' +
str(no_metro_beta))
mean_congested_time = m.mean_edge_attr_per(
layers=['streets'],
attr='congested_time_m_' + str(no_metro_beta),
weight_attr='flow_' + str(no_metro_beta))
dist = m.mean_edge_attr_per(layers=['streets'],
attr='dist_km',
weight_attr='flow_' + str(no_metro_beta))
v_f = dist * 1.0 / mean_free_flow_time
v_c = dist * 1.0 / mean_congested_time
print 'Mean free flow speed = ' + str(round(v_f, 2)) + 'km per minute'
print 'Mean congested speed = ' + str(round(v_c, 2)) + 'km per minute'
# Scale metro so that it runs at v_c (this is beta = 1)
mean_metro_time = m.mean_edge_attr_per(layers=['metro'],
attr='free_flow_time_m')
mean_metro_dist = m.mean_edge_attr_per(layers=['metro'], attr='dist_km')
v_m = mean_metro_dist / mean_metro_time
print 'Mean metro speed = ' + str(round(v_m, 2)) + 'km per minute'
ratio = v_m / v_c
print 'Scaling metro speed by factor of ' + str(round(1 / ratio, 2))
m.scale_edge_attribute(layer='metro',
attribute='free_flow_time_m',
beta=ratio)
print 'Metro now runs at mean speed of street layer'
# run
for beta in betas:
m.scale_edge_attribute(layer='metro',
attribute='free_flow_time_m',
beta=beta)
simulate(m, beta=beta, n=50000)
m.scale_edge_attribute(layer='metro',
attribute='free_flow_time_m',
beta=1.0 / beta)