def test_fitted_value_iteration_pendulum(self):
plant = PendulumPlant()
simulator = Simulator(plant)
def quadratic_regulator_cost(context):
x = context.get_continuous_state_vector().CopyToVector()
x[0] = x[0] - math.pi
u = plant.EvalVectorInput(context, 0).CopyToVector()
return x.dot(x) + u.dot(u)
qbins = np.linspace(0., 2. * math.pi, 51)
qdotbins = np.linspace(-10., 10., 41)
state_grid = [set(qbins), set(qdotbins)]
input_limit = 2.
input_mesh = [set(np.linspace(-input_limit, input_limit, 9))]
timestep = 0.01
[Q, Qdot] = np.meshgrid(qbins, qdotbins)
fig = plt.figure()
ax = fig.gca(projection='3d')
def draw(iteration, mesh, cost_to_go, policy):
# Drawing is slow, don't draw every frame.
if iteration % 20 != 0:
return
plt.title("iteration " + str(iteration))
J = np.reshape(cost_to_go, Q.shape)
surf = ax.plot_surface(Q,
Qdot,
J,
rstride=1,
cstride=1,
cmap=color_map.jet)
plt.pause(0.00001)
surf.remove()
options = DynamicProgrammingOptions()
options.convergence_tol = 1.
options.state_indices_with_periodic_boundary_conditions = {0}
options.visualization_callback = draw
policy, cost_to_go = FittedValueIteration(simulator,
quadratic_regulator_cost,
state_grid, input_mesh,
timestep, options)
def test_fitted_value_iteration_pendulum(self):
plant = PendulumPlant()
simulator = Simulator(plant)
def quadratic_regulator_cost(context):
x = context.get_continuous_state_vector().CopyToVector()
x[0] = x[0] - math.pi
u = plant.EvalVectorInput(context, 0).CopyToVector()
return x.dot(x) + u.dot(u)
qbins = np.linspace(0., 2.*math.pi, 51)
qdotbins = np.linspace(-10., 10., 41)
state_grid = [set(qbins), set(qdotbins)]
input_limit = 2.
input_mesh = [set(np.linspace(-input_limit, input_limit, 9))]
timestep = 0.01
[Q, Qdot] = np.meshgrid(qbins, qdotbins)
fig = plt.figure()
ax = fig.gca(projection='3d')
def draw(iteration, mesh, cost_to_go, policy):
# Drawing is slow, don't draw every frame.
if iteration % 20 != 0:
return
plt.title("iteration " + str(iteration))
J = np.reshape(cost_to_go, Q.shape)
surf = ax.plot_surface(Q, Qdot, J, rstride=1, cstride=1,
cmap=color_map.jet)
plt.pause(0.00001)
surf.remove()
options = DynamicProgrammingOptions()
options.convergence_tol = 1.
options.state_indices_with_periodic_boundary_conditions = {0}
options.visualization_callback = draw
policy, cost_to_go = FittedValueIteration(simulator,
quadratic_regulator_cost,
state_grid, input_mesh,
timestep, options)
return 2 * x.dot(x) + u.dot(u)
if (False):
cost_function = min_time_cost
input_limit = 1.
options.convergence_tol = 0.001
else:
cost_function = quadratic_regulator_cost
input_limit = 3.
options.convergence_tol = 0.1
qbins = np.linspace(0., 2. * math.pi, 51)
qdotbins = np.linspace(-10., 10., 51)
state_grid = [set(qbins), set(qdotbins)]
options.state_indices_with_periodic_boundary_conditions = {0}
input_mesh = [set(np.linspace(-input_limit, input_limit, 9))]
timestep = 0.01
[Q, Qdot] = np.meshgrid(qbins, qdotbins)
fig = plt.figure()
ax = fig.gca(projection='3d')
ax.set_xlabel("theta")
ax.set_ylabel("thetadot")
fig2 = plt.figure()
ax2 = fig2.gca(projection='3d')
ax2.set_xlabel("q")
ax2.set_ylabel("qdot")