Driver Debug Printing
Contents
Driver Debug Printing¶
When working with a model, it may sometimes be helpful to print out the design variables, constraints, and objectives as the Driver
iterates. OpenMDAO provides options on the Driver to let you do that.
Driver Options¶
Option | Default | Acceptable Values | Acceptable Types | Description |
---|---|---|---|---|
debug_print | [] | ['desvars', 'nl_cons', 'ln_cons', 'objs', 'totals'] | ['list'] | List of what type of Driver variables to print at each iteration. |
Usage¶
This example shows how to use the Driver
debug printing options. The debug_print
option is a list of strings. Valid strings include ‘desvars’, ‘ln_cons’, ‘nl_cons’, and ‘objs’. Note that the values for the design variables printed are unscaled, physical values.
Paraboloid
class definition
class Paraboloid(om.ExplicitComponent):
"""
Evaluates the equation f(x,y) = (x-3)^2 + xy + (y+4)^2 - 3.
"""
def setup(self):
self.add_input('x', val=0.0)
self.add_input('y', val=0.0)
self.add_output('f_xy', val=0.0)
def setup_partials(self):
self.declare_partials('*', '*')
def compute(self, inputs, outputs):
"""
f(x,y) = (x-3)^2 + xy + (y+4)^2 - 3
Optimal solution (minimum): x = 6.6667; y = -7.3333
"""
x = inputs['x']
y = inputs['y']
outputs['f_xy'] = (x-3.0)**2 + x*y + (y+4.0)**2 - 3.0
def compute_partials(self, inputs, partials):
"""
Jacobian for our paraboloid.
"""
x = inputs['x']
y = inputs['y']
partials['f_xy', 'x'] = 2.0*x - 6.0 + y
partials['f_xy', 'y'] = 2.0*y + 8.0 + x
import openmdao.api as om
from openmdao.test_suite.components.paraboloid import Paraboloid
prob = om.Problem()
model = prob.model
model.add_subsystem('comp', Paraboloid(), promotes=['*'])
model.add_subsystem('con', om.ExecComp('c = - x + y'), promotes=['*'])
model.set_input_defaults('x', 50.0)
model.set_input_defaults('y', 50.0)
prob.set_solver_print(level=0)
prob.driver = om.ScipyOptimizeDriver()
prob.driver.options['optimizer'] = 'SLSQP'
prob.driver.options['tol'] = 1e-9
prob.driver.options['disp'] = False
prob.driver.options['debug_print'] = ['desvars','ln_cons','nl_cons','objs']
model.add_design_var('x', lower=-50.0, upper=50.0)
model.add_design_var('y', lower=-50.0, upper=50.0)
model.add_objective('f_xy')
model.add_constraint('c', upper=-15.0)
prob.setup()
prob.run_driver()
Driver debug print for iter coord: rank0:ScipyOptimize_SLSQP|0
--------------------------------------------------------------
Design Vars
{'x': array([50.]), 'y': array([50.])}
Nonlinear constraints
{'con.c': array([0.])}
Linear constraints
None
Objectives
{'comp.f_xy': array([7622.])}
Driver debug print for iter coord: rank0:ScipyOptimize_SLSQP|1
--------------------------------------------------------------
Design Vars
{'x': array([50.]), 'y': array([50.])}
Nonlinear constraints
{'con.c': array([0.])}
Linear constraints
None
Objectives
{'comp.f_xy': array([7622.])}
Driver debug print for iter coord: rank0:ScipyOptimize_SLSQP|2
--------------------------------------------------------------
Design Vars
{'x': array([-35.]), 'y': array([-50.])}
Nonlinear constraints
{'con.c': array([-15.])}
Linear constraints
None
Objectives
{'comp.f_xy': array([5307.])}
Driver debug print for iter coord: rank0:ScipyOptimize_SLSQP|3
--------------------------------------------------------------
Design Vars
{'x': array([7.16706813]), 'y': array([-7.83293187])}
Nonlinear constraints
{'con.c': array([-15.])}
Linear constraints
None
Objectives
{'comp.f_xy': array([-27.08333285])}
Driver debug print for iter coord: rank0:ScipyOptimize_SLSQP|4
--------------------------------------------------------------
Design Vars
{'x': array([7.16666667]), 'y': array([-7.83333333])}
Nonlinear constraints
{'con.c': array([-15.])}
Linear constraints
None
Objectives
{'comp.f_xy': array([-27.08333333])}
False
We can also use the debug printing to print some basic information about the derivative calculations so that you can see which derivative is being solved, how long it takes, and the computed values by including the ‘totals’ string in the “debug_print” list.
import openmdao.api as om
from openmdao.test_suite.components.paraboloid import Paraboloid
prob = om.Problem()
model = prob.model
model.add_subsystem('comp', Paraboloid(), promotes=['*'])
model.add_subsystem('con', om.ExecComp('c = - x + y'), promotes=['*'])
model.set_input_defaults('x', 50.0)
model.set_input_defaults('y', 50.0)
prob.set_solver_print(level=0)
prob.driver = om.ScipyOptimizeDriver()
prob.driver.options['optimizer'] = 'SLSQP'
prob.driver.options['tol'] = 1e-9
prob.driver.options['disp'] = False
prob.driver.options['debug_print'] = ['totals']
model.add_design_var('x', lower=-50.0, upper=50.0)
model.add_design_var('y', lower=-50.0, upper=50.0)
model.add_objective('f_xy')
model.add_constraint('c', upper=-15.0)
prob.setup()
prob.run_driver()
Driver total derivatives for iteration: 2
-----------------------------------------
In mode: fwd.
('x', [0])
Elapsed Time: 0.0004057884216308594 secs
In mode: fwd.
('y', [1])
Elapsed Time: 0.0004343986511230469 secs
{('comp.f_xy', 'x'): array([[144.]])}
{('comp.f_xy', 'y'): array([[158.]])}
{('con.c', 'x'): array([[-1.]])}
{('con.c', 'y'): array([[1.]])}
Driver total derivatives for iteration: 3
-----------------------------------------
In mode: fwd.
('x', [0])
Elapsed Time: 0.0002918243408203125 secs
In mode: fwd.
('y', [1])
Elapsed Time: 0.0004260540008544922 secs
{('comp.f_xy', 'x'): array([[-126.]])}
{('comp.f_xy', 'y'): array([[-127.]])}
{('con.c', 'x'): array([[-1.]])}
{('con.c', 'y'): array([[1.]])}
Driver total derivatives for iteration: 4
-----------------------------------------
In mode: fwd.
('x', [0])
Elapsed Time: 0.000537872314453125 secs
In mode: fwd.
('y', [1])
Elapsed Time: 0.00043487548828125 secs
{('comp.f_xy', 'x'): array([[0.50120438]])}
{('comp.f_xy', 'y'): array([[-0.49879562]])}
{('con.c', 'x'): array([[-1.]])}
{('con.c', 'y'): array([[1.]])}
Driver total derivatives for iteration: 5
-----------------------------------------
In mode: fwd.
('x', [0])
Elapsed Time: 0.0003457069396972656 secs
In mode: fwd.
('y', [1])
Elapsed Time: 0.00027680397033691406 secs
{('comp.f_xy', 'x'): array([[0.5]])}
{('comp.f_xy', 'y'): array([[-0.5]])}
{('con.c', 'x'): array([[-1.]])}
{('con.c', 'y'): array([[1.]])}
False