def setUp(self):
"""Setup simple problem"""
np.random.seed(1)
# This needs to work for different
p = 50
n = 200
F = np.random.randn(n, p)
D = np.diag(np.random.rand(n) * np.sqrt(p))
Sigma = F.dot(F.T) + D
gamma = 1.0
mu = cp.Parameter(n, name='mu')
x = cp.Variable(n)
cost = -mu @ x + gamma * cp.quad_form(x, Sigma) + .5 * cp.norm(x, 1)
constraints = [cp.sum(x) == 1, x >= 0]
problem = cp.Problem(cp.Minimize(cost), constraints)
# Define optimizer
# Force mosek to be single threaded
m = mlopt.Optimizer(problem,
# log_level=logging.DEBUG,
)
'''
Sample points
'''
theta_bar = 10 * np.random.rand(n)
radius = 0.8
'''
Train and solve
'''
# Training and testing data
n_train = 100
n_test = 10
# Sample points from multivariate ball
X_d = uniform_sphere_sample(theta_bar, radius, n=n_train)
self.df_train = pd.DataFrame({'mu': list(X_d)})
# # Train and test using pytorch
# params = {
# 'learning_rate': [0.01],
# 'batch_size': [100],
# 'n_epochs': [200]
# }
#
# m.train(df, parallel=False, learner=mlopt.PYTORCH, params=params)
# Testing data
X_d_test = uniform_sphere_sample(theta_bar, radius, n=n_test)
df_test = pd.DataFrame({'mu': list(X_d_test)})
# Store stuff
self.m = m
self.df_test = df_test
def setUp(self):
"""Setup simple problem"""
np.random.seed(1)
# This needs to work for different
p = 10
n = 30
F = np.random.randn(n, p)
D = np.diag(np.random.rand(n) * np.sqrt(p))
Sigma = F.dot(F.T) + D
gamma = 1.0
mu = cp.Parameter(n, name='mu')
x = cp.Variable(n)
cost = -mu @ x + gamma * cp.quad_form(x, Sigma) + .5 * cp.norm(x, 1)
constraints = [cp.sum(x) == 1, x >= 0]
problem = cp.Problem(cp.Minimize(cost), constraints)
# Define optimizer
m = mlopt.Optimizer(problem,
# log_level=logging.DEBUG
)
'''
Sample points
'''
theta_bar = 10 * np.random.rand(n)
radius = 1.0
'''
Train and solve
'''
# Training and testing data
n_train = 300
n_test = 10
# Sample points from multivariate ball
X_d = uniform_sphere_sample(theta_bar, radius, n=n_train)
df = pd.DataFrame({'mu': list(X_d)})
m.train(df,
filter_strategies=True,
parallel=False,
learner=mlopt.PYTORCH,
n_train_trials=10)
# m.train(df, parallel=False, learner=mlopt.XGBOOST, n_train_trials=10)
# Testing data
X_d_test = uniform_sphere_sample(theta_bar, radius, n=n_test)
df_test = pd.DataFrame({'mu': list(X_d_test)})
# Store stuff
self.m = m
self.df_test = df_test
def test_parallel_resolve(self):
"""Test parallel resolve (to avoid hanging)"""
np.random.seed(1)
# This needs to work for different
p = 10
n = p * 10
F = np.random.randn(n, p)
D = np.diag(np.random.rand(n) * np.sqrt(p))
Sigma = F.dot(F.T) + D
gamma = 1.0
mu = cp.Parameter(n, name='mu')
x = cp.Variable(n)
cost = -mu @ x + gamma * cp.quad_form(x, Sigma)
constraints = [cp.sum(x) == 1, x >= 0]
# Define optimizer
problem = cp.Problem(cp.Minimize(cost), constraints)
m = Optimizer(problem, name="portfolio")
'''
Sample points
'''
theta_bar = np.random.randn(n)
radius = 1.0
'''
Train and solve
'''
# Training and testing data
n_train = 1000
n_test = 1000
# Sample points from multivariate ball
X_d = uniform_sphere_sample(theta_bar, radius, n=n_train)
X_d_test = uniform_sphere_sample(theta_bar, radius, n=n_test)
df = pd.DataFrame({'mu': list(X_d)})
df_test = pd.DataFrame({'mu': list(X_d_test)})
# Train and test using pytorch
m.train(df,
parallel=True,
filter_strategies=True,
n_train_trials=10,
learner=PYTORCH)
m.performance(df_test, parallel=True)
# Run parallel loop again to enforce instability
# in multiprocessing
m.performance(df_test, parallel=True)
def test_parallel_vs_serial_learning(self):
"""Test parallel VS serial learning"""
# Generate data
np.random.seed(1)
T = 5
M = 2.
h = 1.
c = 1.
p = 1.
x_init = 2.
radius = 2.
n_train = 1000 # Number of samples
# Define problem
x = cp.Variable(T + 1)
u = cp.Variable(T)
# Define parameter and sampling points
d = cp.Parameter(T, nonneg=True, name="d")
d_bar = 3. * np.ones(T)
X_d = uniform_sphere_sample(d_bar, radius, n=n_train)
df = pd.DataFrame({'d': list(X_d)})
# Constaints
constraints = [x[0] == x_init]
for t in range(T):
constraints += [x[t + 1] == x[t] + u[t] - d[t]]
constraints += [u >= 0, u <= M]
# Objective
cost = cp.sum(cp.maximum(h * x, -p * x)) + c * cp.sum(u)
# Define problem
cvxpy_problem = cp.Problem(cp.Minimize(cost), constraints)
problem = Problem(cvxpy_problem)
# Solve for all theta in serial
results_serial = problem.solve_parametric(df, parallel=False)
# Solve for all theta in parallel
results_parallel = problem.solve_parametric(df, parallel=True)
# Assert all results match
for i in range(n_train):
serial = results_serial[i]
parallel = results_parallel[i]
# Compare x
npt.assert_array_almost_equal(serial['x'],
parallel['x'],
decimal=TOL)
# Compare cost
npt.assert_array_almost_equal(serial['cost'],
parallel['cost'],
decimal=TOL)
# Compare strategy
self.assertTrue(serial['strategy'] == parallel['strategy'])
def sampling_function(n):
"""
Sample data points.
"""
theta_bar = 3. * np.ones(5)
# Sample points from multivariate ball
X = uniform_sphere_sample(theta_bar, 1, n=n)
df = pd.DataFrame({'d': list(X)})
return df
def sample_portfolio(n, k, T=5, N=100):
"""Sample portfolio parameters."""
rad = 0.01
# mean values
np.random.seed(0)
SigmaF_FT_bar = np.random.randn(k, n)
sqrt_D_bar = np.random.rand(n)
# Sigma_F_diag_bar = np.random.rand(k)
hat_r_bar = np.random.rand(n)
w_init_bar = np.random.rand(n)
w_init_bar /= np.sum(w_init_bar)
df = pd.DataFrame()
for i in range(N):
w_init = uniform_sphere_sample(w_init_bar, rad).flatten()
w_init /= np.sum(w_init)
SigmaF_FT = uniform_sphere_sample(SigmaF_FT_bar.flatten(),
rad).reshape(k, n)
x = pd.Series({
# "F": F,
"sqrt_D":
uniform_sphere_sample(sqrt_D_bar, rad).flatten(),
"SigmaF_FT":
SigmaF_FT,
"w_init":
w_init,
})
for t in range(1, T + 1):
x["hat_r_%s" % str(t)] = uniform_sphere_sample(hat_r_bar,
rad).flatten()
df = df.append(x, ignore_index=True)
return df
def setUp(self):
# Generate data
np.random.seed(0)
T = 5
M = 2.
h = 1.
c = 1.
p = 1.
x_init = 2.
self.radius = 2.
n = 1000 # Number of points
n_test = 100
# Define problem
x = cp.Variable(T + 1)
u = cp.Variable(T)
# Define parameter and sampling points
d = cp.Parameter(T, nonneg=True, name="d")
self.d_bar = 2. * np.ones(T)
X_d = uniform_sphere_sample(self.d_bar, self.radius, n=n)
X_d_test = uniform_sphere_sample(self.d_bar, self.radius, n=n_test)
self.df = pd.DataFrame({'d': list(X_d)})
self.df_test = pd.DataFrame({'d': list(X_d_test)})
# Constaints
constraints = [x[0] == x_init]
for t in range(T):
constraints += [x[t + 1] == x[t] + u[t] - d[t]]
constraints += [u >= 0, u <= M]
self.constraints = constraints
# Objective
self.cost = cp.sum(cp.maximum(
h * x, -p * x)) + c * cp.sum_squares(u) + 0.001 * cp.sum_squares(x)
# Define problem
problem = cp.Problem(cp.Minimize(self.cost), self.constraints)
self.optimizer = Optimizer(problem)
