def experiment_driving_dist(ev,
CS_density,
min_dist,
max_dist,
step_size,
iterations,
file_name='driving_dist.csv'):
print('loading road network...')
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
print('setting charge stations...')
charge_station_density(rn, CS_density)
with open(file_name, 'a') as f:
f.write(
'driving distance, naive-time, naive-fail, greedy-time, greedy-fail\n'
)
for distance in range(min_dist, max_dist, step_size):
print 'driving distance experiment. currently at: ', distance
naive_t, greedy_t = 0, 0
naive_f, greedy_f = 0, 0
for iteration in xrange(iterations):
s, t, dist = s_and_t(rn, distance)
print 'iteration: ', iteration + 1, ' distance: ', dist
print 's: ', rn.edges([s],
data=True)[0], ' t: ', rn.edges([t],
data=True)[0]
### NAIVE
print 'Initialising Naive...'
_, time = naive_path(rn, s, t, ev)
naive_t += time if time != float('inf') else 0
naive_f += 0 if time != float('inf') else 1
linear_t1, cur = linearProgramming(rn, _, ev, ev.curbat)
print 'done'
### Greedy
print 'Initialising Greedy...'
_, time = fastest_path_greedy(rn, s, t, 1, ev) # 1 for slope
greedy_t += time if time != float('inf') else 0
greedy_f += 0 if time != float('inf') else 1
linear_t2, cur = linearProgramming(rn, _, ev, ev.curbat)
print naive_t, linear_t1, greedy_t, linear_t2
print 'done'
with open(file_name, 'a') as f:
f.write('%s,%s,%s,%s,%s\n' % (
distance,
naive_t /
(iterations - naive_f) if iterations != naive_f else 'inf',
naive_f,
greedy_t /
(iterations - greedy_f) if iterations != greedy_f else 'inf',
greedy_f,
))
def experiment_runtime_compexity(ev,
iterations,
path_distance,
CS_density,
step_size,
file_name='time_complexity.csv'):
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
charge_station_density(rn, CS_density)
number_of_nodes = 400000
with open(file_name, 'a') as f:
f.write('nodes,dijkstra,greedy\n')
while number_of_nodes > 0:
actual_nodes = set_roadnetwork_complexity(rn, number_of_nodes)
dijkstra_t, greedy_t = 0, 0
for x in range(iterations):
s, t, dist = s_and_t(rn, path_distance)
### Dijkstras
start_time = time.time()
nx.single_source_dijkstra_path_length(rn, s, weight='t')
dijkstra_t += time.time() - start_time
### Greedy
start_time = time.time()
fastest_path_greedy(rn, s, t, 1, ev)
greedy_t += time.time() - start_time
with open(file_name, 'a') as f:
f.write(
'%s,%s\n' %
(actual_nodes, dijkstra_t / iterations, greedy_t / iterations))
number_of_nodes -= step_size
def experiment_charge_rate(ev,
iterations,
charge_rate_variance,
path_distance,
CS_density,
step_size,
file_name='charge_rate.csv'):
with open(file_name, 'a') as f:
f.write(
'Charge rate scale,naive-time,naive-fail,greedy-time,greedy-fail\n'
)
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
s, t, dist = s_and_t(rn, path_distance)
for scale_factor in range(-charge_rate_variance, charge_rate_variance,
step_size):
print 'charge station experiment. currently at: ', 1 + (scale_factor /
100.0)
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
set_roadnetwork_complexity(rn, 500000)
charge_station_density(rn, CS_density)
scale_charge_rates(rn, 1 + (scale_factor / 100.0))
naive_t, greedy_t = 0, 0
naive_f, greedy_f = 0, 0
for iteration in range(iterations):
### NAIVE
_, time = naive_path(rn, s, t, ev)
naive_t += time if time != float('inf') else 0
naive_f += 0 if time != float('inf') else 1
### Greedy
_, time = fastest_path_greedy(rn, s, t, 1, ev) # 1 for slope
greedy_t += time if time != float('inf') else 0
greedy_f += 0 if time != float('inf') else 1
with open(file_name, 'a') as f:
f.write('%s,%s,%s,%s,%s,%s,%s\n' % (
scale_factor,
naive_t /
(iterations - naive_f) if iterations != naive_f else 'inf',
naive_f,
greedy_t /
(iterations - greedy_f) if iterations != greedy_f else 'inf',
greedy_f,
))
def experiment_cs_density(ev,
iterations,
path_distance,
file_name='cs_density.csv'):
with open(file_name, 'a') as f:
f.write('CS density,naive-time,naive-fail,greedy-time,greedy-fail\n')
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
s, t, dist = s_and_t(rn, path_distance)
for cs_dist in range(8, 55, 3):
print 'CS density experiment. currently at: ', cs_dist
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
charge_station_density(rn, cs_dist)
naive_t, greedy_t = 0, 0
naive_f, greedy_f = 0, 0
for iteration in range(iterations):
### NAIVE
_, time = naive_path(rn, s, t, ev)
naive_t += time if time != float('inf') else 0
naive_f += 0 if time != float('inf') else 1
### Greedy
_, time = fastest_path_greedy(rn, s, t, 1, ev) # 1 for slope
greedy_t += time if time != float('inf') else 0
greedy_f += 0 if time != float('inf') else 1
with open(file_name, 'a') as f:
f.write('%s,%s,%s,%s,%s\n' % (
cs_dist,
naive_t /
(iterations - naive_f) if iterations != naive_f else 'inf',
naive_f,
greedy_t /
(iterations - greedy_f) if iterations != greedy_f else 'inf',
greedy_f,
))
subprocess.call([
"say",
str(greedy_t /
(iterations - greedy_f)) if iterations != greedy_f else 'inf'
])
def experiment_ev_consumption(iterations,
path_distance,
con_rate_variance,
CS_density,
file_name='ev_consumption.csv'):
rn = RoadNetwork(nx.read_gpickle('pickle_experiment'))
charge_station_density(rn, cs_dist)
s, t, dist = s_and_t(rn, path_distance)
with open(file_name, 'a') as f:
f.write('CS density,naive-time,naive-fail,greedy-time,greedy-fail\n')
ev_number = 0
for ev in scale_cons_rate(con_rate_variance): #Create vehicles
print 'charge station experiment. currently at EV: ', ev_number
scale_charge_rates(rn, 1 + (scale_factor / 100.0))
naive_t, greedy_t = 0, 0
naive_f, greedy_f = 0, 0
for iteration in range(iterations):
### NAIVE
_, time = naive_path(rn, s, t, ev)
naive_t += time if time != float('inf') else 0
naive_f += 0 if time != float('inf') else 1
### Greedy
_, time = fastest_path_greedy(rn, s, t, 1, ev) # 1 for slope
greedy_t += time if time != float('inf') else 0
greedy_f += 0 if time != float('inf') else 1
with open(file_name, 'a') as f:
f.write('%s,%s,%s,%s,%s\n' % (
scale_factor,
naive_t /
(iterations - naive_f) if iterations != naive_f else 'inf',
naive_f,
greedy_t /
(iterations - greedy_f) if iterations != greedy_f else 'inf',
greedy_f,
))
ev_number += 1