def test_solver(self):
# Frank-Wolfe from Algorithm 1 of Jaggi's paper
print 'test solver'
graph = np.loadtxt('data/braess_net.csv', delimiter=',', skiprows=1)
demand = np.loadtxt('data/braess_od.csv', delimiter=',', skiprows=1)
demand=np.reshape(demand, (1,3))
f = solver(graph, demand, max_iter=300)
self.check(f, np.array([1.,1.,0.,1.,1.]), 1e-2)
# modify demand
demand[0,2] = 0.5
f = solver(graph, demand)
self.check(f, np.array([.5,.0,.5,.0,.5]), 1e-8)
def test_solver_sioux_falls(self):
print 'test Frank-Wolfe on Sioux Falls'
graph = np.loadtxt('data/SiouxFalls_net.csv', delimiter=',', skiprows=1)
demand = np.loadtxt('data/SiouxFalls_od.csv', delimiter=',', skiprows=1)
demand[:,2] = demand[:,2] / 4000
f = solver(graph, demand, max_iter=1000)
results = np.loadtxt('data/SiouxFalls_results.csv')
self.check(f*4000, results, 1e-3)
def test_anaheim(self):
print 'test Frank-Wolfe on Anaheim'
graph = np.loadtxt('data/Anaheim_net.csv', delimiter=',', skiprows=1)
demand = np.loadtxt('data/Anaheim_od.csv', delimiter=',', skiprows=1)
demand[:, 2] = demand[:, 2] / 4000
f = solver(graph, demand, max_iter=1000)
# print f.shape
results = np.loadtxt('data/Anaheim_results.csv')
print np.linalg.norm(f * 4000 - results) / np.linalg.norm(results)
def Cython_Func_LA():
#graph = np.loadtxt('data/SiouxFalls_net.csv', delimiter=',', skiprows=1)
#demand = np.loadtxt('data/SiouxFalls_od.csv', delimiter=',', skiprows=1)
graph = np.loadtxt('data/LA_net.csv', delimiter=',', skiprows=1)
demand = np.loadtxt('data/LA_od_2.csv', delimiter=',', skiprows=1)
graph[10787, -1] = graph[10787, -1] / (1.5**4)
graph[3348, -1] = graph[3348, -1] / (1.2**4)
demand[:, 2] = 0.5 * demand[:, 2] / 4000
#import pdb; pdb.set_trace()
f = solver(graph, demand, max_iter=30)
#results = np.loadtxt('data/SiouxFalls_results.csv')
np.savetxt('data/la/LA_Cython.csv', f, delimiter=',')
def frank_wolfe_on_chicago():
"""
Frank-Wolfe on Chicago with the inputs processed from:
http://www.bgu.ac.il/~bargera/tntp/
"""
graph, demand, node, features = load_chicago()
results = np.loadtxt("data/Chicago_results.csv", delimiter=",", skiprows=1)
demand[:, 2] = demand[:, 2] / 4000
f = solver(graph, demand, max_iter=1000, display=1)
# error: 0.00647753330249, time: 664.151s
# f = solver_2(graph, demand, max_iter=1000, q=100, display=1)
# error: 0.00646125552755, time: 664.678s
# f = solver_3(graph, demand, max_iter=1000, q=200, display=1)
# error: 0.00648532089623, time: 665.074s
print np.linalg.norm(f * 4000 - results[:, 2]) / np.linalg.norm(results[:, 2])
def frank_wolfe_on_chicago():
'''
Frank-Wolfe on Chicago with the inputs processed from:
http://www.bgu.ac.il/~bargera/tntp/
'''
graph, demand, node, features = load_chicago()
results = np.loadtxt('data/Chicago_results.csv', delimiter=',', skiprows=1)
demand[:,2] = demand[:,2] / 4000
f = solver(graph, demand, max_iter=1000, display=1)
# error: 0.00647753330249, time: 664.151s
# f = solver_2(graph, demand, max_iter=1000, q=100, display=1)
# error: 0.00646125552755, time: 664.678s
# f = solver_3(graph, demand, max_iter=1000, q=200, display=1)
# error: 0.00648532089623, time: 665.074s
print np.linalg.norm(f*4000 - results[:,2]) / np.linalg.norm(results[:,2])
def gauss_seidel(graphs, demands, solver, max_cycles=10, max_iter=100, \
by_origin=False, q=10, display=0, past=10, stop=1e-8, eps=1e-8, \
stop_cycle=None):
# we are given a list of graphs and demands that are specific for different types of players
# the gauss-seidel scheme updates cyclically for each type at a time
if stop_cycle is None:
stop_cycle = stop
g = construct_igraph(graphs[0])
ods = [construct_od(d) for d in demands]
types = len(graphs)
fs = np.zeros((graphs[0].shape[0],types),dtype="float64")
g2 = np.copy(graphs[0])
K = total_ff_costs_heterogeneous(graphs, g, ods)
if K < eps:
K = np.sum([np.sum(d[:,2]) for d in demands])
elif display >= 1:
print 'average free-flow travel time', \
K / np.sum([np.sum(d[:,2]) for d in demands])
for cycle in range(max_cycles):
if display >= 1: print 'cycle:', cycle
for i in range(types):
# construct graph with updated latencies
shift = np.sum(fs[:,range(i)+range(i+1,types)], axis=1)
shift_graph(graphs[i], g2, shift)
g.es["weight"] = g2[:,3].tolist()
# update flow assignment for this type
fs[:,i] = solver(g2, demands[i], g, ods[i], max_iter=max_iter, q=q, \
display=display, past=past, stop=stop, eps=eps)
# check if we have convergence
r = residual(graphs, demands, g, ods, fs) / K
if display >= 1:
print 'error:', r
if r < stop_cycle and r > 0:
return fs
return fs
