"""
This file contains functions needed to run Aviary using the Level 1 interface.
"""
import os
from importlib.machinery import SourceFileLoader
from pathlib import Path
import openmdao.api as om
from aviary.variable_info.enums import AnalysisScheme, Verbosity
from aviary.interface.methods_for_level2 import AviaryProblem
from aviary.utils.functions import get_path
[docs]
def run_aviary(aircraft_filename, phase_info, optimizer=None,
analysis_scheme=AnalysisScheme.COLLOCATION, objective_type=None,
record_filename='problem_history.db', restart_filename=None, max_iter=50,
run_driver=True, make_plots=True, phase_info_parameterization=None,
optimization_history_filename=None, verbosity=Verbosity.BRIEF):
"""
Run the Aviary optimization problem for a specified aircraft configuration and mission.
This function creates an instance of the AviaryProblem class using provided phase information,
mission, and mass methods. It processes aircraft and options data from the given aircraft filename,
checks for clashing inputs, and sets up pre- and post-mission systems. The optimization problem is formulated,
initialized with guesses, and run to find a solution.
Parameters
----------
aircraft_filename : str
Filename from which to load the aircraft and options data.
phase_info : dict
Information about the phases of the mission.
optimizer : str
The optimizer to use.
analysis_scheme : AnalysisScheme, optional
The analysis scheme to use, defaults to AnalysisScheme.COLLOCATION.
objective_type : str, optional
Type of the optimization objective.
record_filename : str, optional
Filename for recording the solution, defaults to 'dymos_solution.db'.
restart_filename : str, optional
Filename to use for restarting the optimization, if applicable.
max_iter : int, optional
Maximum number of iterations for the optimizer, defaults to 50.
run_driver : bool, optional
If True, the driver will be run, defaults to True.
make_plots : bool, optional
If True, generate plots during the optimization, defaults to True.
phase_info_parameterization : function, optional
Additional information to parameterize the phase_info object based on
desired cruise altitude and Mach.
optimization_history_filename : str or Path
The name of the database file where the driver iterations are to be recorded. The
default is None.
verbosity : Verbosity or int
Sets level of information outputted to the terminal during model execution.
Returns
-------
AviaryProblem
The AviaryProblem instance after running the optimization problem.
Notes
-----
The function allows for user overrides on aircraft and options data.
It raises warnings or errors if there are clashing user inputs.
Users can modify or add methods to alter the Aviary problem's behavior.
"""
# compatibility with being passed int for verbosity
verbosity = Verbosity(verbosity)
# Build problem
prob = AviaryProblem(analysis_scheme, name=Path(aircraft_filename).stem)
# Load aircraft and options data from user
# Allow for user overrides here
prob.load_inputs(aircraft_filename, phase_info, verbosity=verbosity)
# Preprocess inputs
prob.check_and_preprocess_inputs()
prob.add_pre_mission_systems()
prob.add_phases(phase_info_parameterization=phase_info_parameterization)
prob.add_post_mission_systems()
# Link phases and variables
prob.link_phases()
prob.add_driver(optimizer, max_iter=max_iter, verbosity=verbosity)
prob.add_design_variables()
# Load optimization problem formulation
# Detail which variables the optimizer can control
prob.add_objective(objective_type=objective_type)
prob.setup()
prob.set_initial_guesses()
prob.run_aviary_problem(
record_filename, restart_filename=restart_filename, run_driver=run_driver, make_plots=make_plots, optimization_history_filename=optimization_history_filename)
return prob
[docs]
def run_level_1(
input_deck,
outdir='output',
optimizer='SNOPT',
phase_info=None,
n2=False,
max_iter=50,
analysis_scheme=AnalysisScheme.COLLOCATION,
):
'''
This file enables running aviary from the command line with a user specified input deck.
usage: aviary run_mission [input_deck] [opt_args]
'''
kwargs = {
'max_iter': max_iter,
}
if analysis_scheme is AnalysisScheme.SHOOTING:
kwargs['analysis_scheme'] = AnalysisScheme.SHOOTING
kwargs['run_driver'] = False
# kwargs['optimizer'] = 'IPOPT'
# else:
kwargs['optimizer'] = optimizer
if isinstance(phase_info, str):
phase_info_path = get_path(phase_info)
phase_info_file = SourceFileLoader(
"phase_info_file", str(phase_info_path)).load_module()
phase_info = getattr(phase_info_file, 'phase_info')
kwargs['phase_info_parameterization'] = getattr(
phase_info_file, 'phase_info_parameterization', None)
prob = run_aviary(input_deck, phase_info, **kwargs)
if n2:
outfile = os.path.join(outdir, "n2.html")
if outdir != '':
os.makedirs(outdir, exist_ok=True)
om.n2(
prob,
outfile=outfile,
show_browser=False,
)
return prob
def _setup_level1_parser(parser):
def_outdir = os.path.join(os.getcwd(), "output")
parser.add_argument(
'input_deck', metavar='indeck', type=str, nargs=1, help='Name of vehicle input deck file'
)
parser.add_argument(
"-o", "--outdir", default=def_outdir, help="Directory to write outputs"
)
parser.add_argument(
"--optimizer",
type=str,
default='SNOPT',
help="Name of optimizer",
choices=("SNOPT", "IPOPT", "SLSQP", "None")
)
parser.add_argument(
"--phase_info",
type=str,
default=None,
help="Path to phase info file"
)
parser.add_argument("--n2", action="store_true",
help="Generate an n2 diagram after the analysis")
parser.add_argument(
"--max_iter",
type=int,
default=50,
help="maximum number of iterations")
parser.add_argument(
"--shooting",
action="store_true",
help="Use shooting instead of collocation",
)
def _exec_level1(args, user_args):
if args.shooting:
analysis_scheme = AnalysisScheme.SHOOTING
else:
analysis_scheme = AnalysisScheme.COLLOCATION
if args.optimizer == 'None':
args.optimizer = None
# check if args.input_deck is a list, if so, use the first element
if isinstance(args.input_deck, list):
args.input_deck = args.input_deck[0]
if args.outdir == os.path.join(os.getcwd(), "output"):
# if default outdir, add the input deck name
file_name_stem = Path(args.input_deck).stem
args.outdir = args.outdir + os.sep + file_name_stem
prob = run_level_1(
input_deck=args.input_deck,
outdir=args.outdir,
optimizer=args.optimizer,
phase_info=args.phase_info,
n2=args.n2,
max_iter=args.max_iter,
analysis_scheme=analysis_scheme,
)
