# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com # # This work is licensed under a BSD 0-Clause License. # # Permission to use, copy, modify, and/or distribute this software # for any purpose with or without fee is hereby granted. # # THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL # WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED # WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL # THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, # OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING # FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, # NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION # WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. # Contributors: # INITIAL AUTHORS - API and implementation and/or documentation # :author: Fabian Castañeda # OTHER AUTHORS - MACROSCOPIC CHANGES """ Bi-level BCD-based MDO on the Sobieski SSBJ test case ===================================================== """ # %% # .. note:: # # As described in :ref:`bcd_formulation`, # there are several variants of the bi-level BCD formulation; this example shows # the implementation of the bi-level BCD-MDF (BL-BCD-MDF). # For more details about other variants please # refer to :cite:`david:hal-04758286`. from __future__ import annotations from copy import deepcopy from gemseo import configure_logger from gemseo import execute_post from gemseo.problems.mdo.sobieski.core.design_space import SobieskiDesignSpace from gemseo.problems.mdo.sobieski.disciplines import SobieskiAerodynamics from gemseo.problems.mdo.sobieski.disciplines import SobieskiMission from gemseo.problems.mdo.sobieski.disciplines import SobieskiPropulsion from gemseo.problems.mdo.sobieski.disciplines import SobieskiStructure from gemseo.scenarios.mdo_scenario import MDOScenario from gemseo.settings.formulations import BiLevel_BCD_Settings from gemseo.settings.formulations import MDF_Settings from gemseo.settings.mda import MDAGaussSeidel_Settings from gemseo.settings.opt import NLOPT_COBYLA_Settings from gemseo.settings.opt import SLSQP_Settings configure_logger() # %% # Instantiate the disciplines # ---------------------------- # First, we instantiate the four disciplines of the use case: # :class:`.SobieskiPropulsion`, # :class:`.SobieskiAerodynamics`, # :class:`.SobieskiMission` # and :class:`.SobieskiStructure`. propulsion_disc = SobieskiPropulsion() aerodynamics_disc = SobieskiAerodynamics() structure_disc = SobieskiStructure() mission_disc = SobieskiMission() # %% # Since they are going to be our disciplines for the sub-scenarios, we'll call them # sub-disciplines. sub_disciplines = [structure_disc, propulsion_disc, aerodynamics_disc, mission_disc] # %% # Build the scenario # ---------------------- # We build the scenario that allows to create the optimization problem from # the disciplines and the formulation. # Here, we use the :class:`.BiLevelBCD` formulation. # We need to define the design space. design_space = SobieskiDesignSpace() # %% # For this formulation, we need to define the optimization sub-scenarios from # all sub-disciplines coupled together. Each sub-scenario optimizes its own design # variable according to the corresponding constraint and the objective y_4 (range) # which we are maximizing. # %% # Define Sub-scenario settings model # -------------------------------------- # The setting for all sub-scenarios is the same, so we can define a global, settings # model to be used by each sub-scenario. sub_scenario_settings = MDF_Settings( main_mda_name="MDAGaussSeidel", ) sc_algo_settings = SLSQP_Settings(max_iter=50) # %% # Build the Propulsion Sub-scenario # ----------------------------------- # This sub-scenario will optimize the propulsion's discipline design variable x_3 under # the constraint g_3. propulsion_sc = MDOScenario( sub_disciplines, "y_4", design_space.filter(["x_3"], copy=True), formulation_settings_model=sub_scenario_settings, maximize_objective=True, name="PropulsionScenario", ) propulsion_sc.set_algorithm(algo_settings_model=sc_algo_settings) propulsion_sc.formulation.optimization_problem.objective *= 0.001 propulsion_sc.add_constraint("g_3", constraint_type="ineq") # %% # Build the Aerodynamics Sub-scenario # ----------------------------------- # This sub-scenario will optimize the aerodynamics' discipline design variable x_2 under # the constraint g_2. aerodynamics_sc = MDOScenario( sub_disciplines, "y_4", design_space.filter(["x_2"], copy=True), formulation_settings_model=sub_scenario_settings, maximize_objective=True, name="AerodynamicsScenario", ) aerodynamics_sc.set_algorithm(algo_settings_model=sc_algo_settings) aerodynamics_sc.formulation.optimization_problem.objective *= 0.001 aerodynamics_sc.add_constraint("g_2", constraint_type="ineq") # %% # Build the Structure Sub-scenario # ----------------------------------- # This sub-scenario will optimize the structure's discipline design variable x_1 under # the constraint g_1. structure_sc = MDOScenario( sub_disciplines, "y_4", design_space.filter(["x_1"], copy=True), formulation_settings_model=sub_scenario_settings, maximize_objective=True, name="StructureScenario", ) structure_sc.set_algorithm(algo_settings_model=sc_algo_settings) structure_sc.formulation.optimization_problem.objective *= 0.001 structure_sc.add_constraint("g_1", constraint_type="ineq") # %% # System's Scenario Settings # --------------------------- # The bi-level BCD formulation allows to independently define the settings # for the BCD MDA, such as shown below. bcd_mda_settings = MDAGaussSeidel_Settings(tolerance=1e-5, max_mda_iter=10) # %% # Then, you may pass the BCD MDA settings directly to the formulation settings. # Since the system constraints are the same as the constraints that have already been # applied to the sub-scenarios, we set ``apply_cstr_tosub_scenarios=False`` to avoid # adding the same constraints twice on the lower level. system_settings = BiLevel_BCD_Settings( bcd_mda_settings=bcd_mda_settings, apply_cstr_tosub_scenarios=False, ) # %% # .. tip:: # # When running bi-level scenarios, it is interesting to access the optimization # history of the sub-scenarios for each system iteration. By default, the setting # ``keep_opt_history`` is set to ``True``. This allows you to store in memory the # databases of the sub-scenarios (see the last section of this example for more # details). # In some cases, storing the databases in memory can take up too much space and cause # performance issues. In these cases, set ``keep_opt_history=False`` and save the # databases to the disk using ``save_opt_history=True``. # %% # Just like for the sub-scenario, we define the algorithm settings for the # system scenario. system_sc_algo_settings = NLOPT_COBYLA_Settings(max_iter=100) # %% # Build the System's Scenario # ----------------------------------- # The system level scenario is based on the three previous sub-scenarios for which we aim to maximize the range. sub_scenarios = [propulsion_sc, aerodynamics_sc, structure_sc, mission_disc] system_scenario = MDOScenario( sub_scenarios, "y_4", design_space.filter(["x_shared"], copy=True), formulation_settings_model=system_settings, maximize_objective=True, ) system_scenario.formulation.optimization_problem.objective *= 0.001 system_scenario.set_algorithm(algo_settings_model=system_sc_algo_settings) system_scenario.add_constraint("g_1", constraint_type="ineq") system_scenario.add_constraint("g_2", constraint_type="ineq") system_scenario.add_constraint("g_3", constraint_type="ineq") # %% # Visualize the XDSM # ^^^^^^^^^^^^^^^^^^ # Generate the XDSM on the fly: # # - ``log_workflow_status=True`` will log the status of the workflow in the console, # - ``save_html`` (default ``True``) will generate a self-contained HTML file, # that can be automatically opened using ``show_html=True``. system_scenario.xdsmize(save_html=False) # %% # Execute the main scenario # ^^^^^^^^^^^^^^^^^^^^^^^^^ system_scenario.execute() # %% # Plot the history of the MDA residuals # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ # For the first MDA: system_scenario.formulation.mda1.plot_residual_history(save=False, show=True) # %% # For the second MDA: system_scenario.formulation.mda2.plot_residual_history(save=False, show=True) # %% # For the BCD MDA: system_scenario.formulation.bcd_mda.plot_residual_history(save=False, show=True) # %% # Plot the system optimization history view # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ system_scenario.post_process(post_name="OptHistoryView", save=False, show=True) # %% # Plot the structure optimization histories of the 2 first iterations # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ struct_databases = system_scenario.formulation.scenario_adapters[2].databases for database in struct_databases[:2]: opt_problem = deepcopy(structure_sc.formulation.optimization_problem) opt_problem.database = database execute_post(opt_problem, post_name="OptHistoryView", save=False, show=True) # %% # Print execution metrics on disciplines and sub-scenarios # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ for disc in [propulsion_disc, aerodynamics_disc, mission_disc, structure_disc]: print(f"{disc.name}: {disc.execution_statistics.n_executions} calls.") for sub_sc in [propulsion_sc, aerodynamics_sc, structure_sc]: print(f"{sub_sc.name}: {sub_sc.execution_statistics.n_executions} calls.")