Multistart optimization

Runs simple optimization problem with multiple starting points Nests a MDOScenario in a DOEScenario using a MDOScenarioAdapter.

from __future__ import absolute_import, division, print_function, unicode_literals

from future import standard_library

from gemseo.api import (
    configure_logger,
    create_design_space,
    create_discipline,
    create_scenario,
)
from gemseo.core.mdo_scenario import MDOScenarioAdapter

configure_logger()

standard_library.install_aliases()

Create the disciplines

objective = create_discipline(
    "AnalyticDiscipline", expressions_dict={"obj": "x**3-x+1"}
)
constraint = create_discipline(
    "AnalyticDiscipline", expressions_dict={"cstr": "x**2+obj**2-1.5"}
)

Create the design space

design_space = create_design_space()
design_space.add_variable("x", 1, l_b=-1.5, u_b=1.5, value=1.5)

Create the MDO scenario

scenario = create_scenario(
    [objective, constraint],
    formulation="DisciplinaryOpt",
    objective_name="obj",
    design_space=design_space,
)
scenario.default_inputs = {"algo": "SLSQP", "max_iter": 10}
scenario.add_constraint("cstr", "ineq")

Create the scenario adapter

dv_names = scenario.formulation.opt_problem.design_space.variables_names
adapter = MDOScenarioAdapter(
    scenario, dv_names, ["obj", "cstr"], set_x0_before_opt=True
)

Create the DOE scenario

scenario_doe = create_scenario(
    adapter,
    formulation="DisciplinaryOpt",
    objective_name="obj",
    design_space=design_space,
    scenario_type="DOE",
)
scenario_doe.add_constraint("cstr", "ineq")
run_inputs = {"n_samples": 10, "algo": "fullfact"}
scenario_doe.execute(run_inputs)

Out:

{'eval_jac': False, 'algo': 'fullfact', 'n_samples': 10}

Plot the optimum objective for different x0

scenario_doe.post_process("BasicHistory", data_list=["obj"], save=False, show=True)

Out:

<gemseo.post.basic_history.BasicHistory object at 0x7fc29d9e7760>