MDO formulation options

List of available MDO formulation: BiLevel - DisciplinaryOpt - IDF - MDF -

BiLevel

Description

A bi-level formulation draws an optimization architecture that involves multiple optimization problems to be solved to obtain the solution of the MDO problem.

Here, at each iteration on the global design variables, the bi-level MDO formulation implementation performs a first MDA to compute the coupling variables, then disciplinary optimizations on the local design variables in parallel and then, a second MDA to update the coupling variables.

Options

  • apply_cstr_to_system, bool - if True, the add_constraint method adds the constraint to the optimization problem of the system scenario.

  • apply_cstr_tosub_scenarios, bool - if True, the add_constraint method adds the constraint to the optimization problem of the sub-scenario capable of computing the constraint.

  • maximize_objective, bool - if True, the objective function is maximized, by default, a minimization is performed.

  • mda_name, str - class name of the MDA to be used.

  • multithread_scenarios, bool - if True and parallel_scenarios=True, the sub scenarios are run in parallel using multi-threading, if False and parallel_scenarios=True, multi-processing is used.

  • parallel_scenarios, bool - if True, the sub scenarios are run in parallel.

  • reset_x0_before_opt, bool - if True, restart the sub optimizations from the initial guesses, otherwise warm start them

DisciplinaryOpt

Description

The disciplinary optimization formulation draws the architecture of a mono disciplinary optimization process from an ordered list of disciplines, an objective function and a design space. The objective function is minimized by default.

Options

  • maximize_objective, bool - if True, the objective function is maximized, by default, a minimization is performed.

IDF

Description

The Individual Discipline Feasible formulation.

This formulation draws an optimization architecture where the coupling variables of strongly coupled disciplines is made consistent by adding equality constraints on the coupling variables at top level, the optimization problem w.r.t. local, global design variables and coupling variables is made at the top level. Disciplinary analysis is made at a each optimization iteration. Multidisciplinary analysis is made at the optimum.

Options

  • maximize_objective, bool - if True, the objective function is maximized, by default, a minimization is performed.

  • normalize_constraints, bool - if True, outputs of the coupling consistency contraints are scaled.

  • parallel_exec, bool - if True, all constraints and objectives are computed in parallel. At every iteration, all disciplines are executed in parallel. Otherwise, a separate constraint is created for each discipline with couplings.

  • start_at_equilibrium, bool - if True, an MDA is used to initialize coupling variables.

  • use_threading, bool - if True and parallel_exec=True, the disciplines are run in parallel using multi-threading, if False and parallel_exec=True, multi-processing is used.

MDF

Description

The Multidisciplinary Design Feasible formulation draws an optimization architecture where the coupling of strongly coupled disciplines is made consistent by means of a Multidisciplinary Design Analysis (MDA), the optimization problem w.r.t. local and global design variables is made at the top level. Multidisciplinary analysis is made at a each optimization iteration.

Options

  • main_mda_class, str - classname of the main MDA, typically the MDAChain, but one can force to use MDAGaussSeidel for instance

  • maximize_objective, bool - if True, the objective function is maximized, by default, a minimization is performed.

  • sub_mda_class, str - the type of MDA to be used, shall be the class name. (default MDAJacobi)