Glossary

AAO

All At Once

algorithm integrator

A user class of the formulation engine that integrates mathematical algorithms such as an optimization algorithm, DOE method or surrogate model

anaconda

A python tool to create virtual environment and easily install precompiled packages, https://www.anaconda.com/distribution

ANN

Artificial Neural Network

ANR

Agence Nationale de la Recherche

API

Application Program Interface

ASO

Asymmetric Subspace Optimization

ATC

Analytical Target Cascading

bi-level

A type of MDO formulation which formulates multiple optimization problems

BLISS

Bi-Level Integrated System Synthesis,

BLISS2000

Bi-Level Integrated System Synthesis, BLISS variant from 2000 paper, using surrogate models

BLISS98

Bi-Level Integrated System Synthesis, original BLISS variant from 1998 paper using coupled derivatives

CFD

Computational Fluid Dynamics

chain

A process that executes a set of disciplines in a sequential way, where outputs of the previously executed disciplines are passed as inputs of the next ones

CO

Collaborative Optimization

complex step

A numerical method to approximate the derivative of a function, similarly to finite differences, but using complex perturbations

constraint

A function of the design variables that must be kept either null or negative in an optimization problem

constraints

All the functions of the design variables that must be kept either null or negative in an optimization problem

COTS

Commercial Off The Shelf

coupling variables

In an MDO scenario, variables that are both used as inputs of a discipline and outputs of another one, or the same discipline

CSM

Computational Structure Mechanics

CSSO

Concurrent Subspaces Optimization

data flow

The sequence of data creation by, and exchanges between disciplines in a process

data-driven

Characteristic of a workflow engine, for which processes are described through the data flow, and in which the work flow is deduced accordingly

design document

A document that describes the design of a program, responding to a Software Requirements Specification

Design Of Experiments

A sampling of a design space, or a generic method that produces samplings of design spaces

design problem

An engineering problem such that a shape has to be changed to match or improve criteria under constraints

design space

The mathematical set containing the design variables of an optimization problem

design variables

The unknowns of the optimization problem

DFO

Derivative-Free Optimization

disciplinary design variables

In an MDO scenario, a subset of the design variables that are inputs of only one discipline

disciplinary expert

A user class of the MDO platform that wraps disciplinary capabilities into disciplines and creates, tests or maintains disciplinary processes

disciplinary optimization

The most basic MDO formulation restricted to a single set of design variables and only suited for a weakly coupled problem

discipline

One program, or an arbitrary set of simulation software, that can be viewed as a mathematical multivalued function, taking inputs and producing outputs through its execution

DOE

Design Of Experiments

DOE algorithm

An algorithm that generates samples of the design space

driver

A optimization algorithm or DOE

EA

Evolutionary Algorithm

eWORMS

eclipse Workflow Management System

finite differences

A numerical method to approximate the derivative of a function by small perturbations of the inputs

fixed point

A family of numerical resolution methods based on an iterative sequence of execution of the type : A() -> B() -> C() -> … -> A() -> B() -> C() until convergence

FPO

Future Projects Office

GEMSEO

Generic Engine for Multi-disciplinary Scenarios, Exploration and Optimization

generic process

A process that can be applied to any discipline or set of disciplines, such as Multi Disciplinary Analyses methods that solves the coupling variables of a set of disciplines

gradient

Total derivative of a function with respect to its variables

gradient-based optimization

A class of optimization algorithms that use the total derivatives of the objective function and constraints

grammar

A set of rules to be respected by a data set. Typically used to describe the inputs and outputs of a discipline

HDF

Hierarchical Data Format to save and structure files with huge data

Hessian

The matrix of second order partial derivatives of one output with respect to inputs

HPC

High Performance Computing

HTML

Hypertext Markup Language

IDF

Individual Discipline Feasible

interface

A set of functions and data from a software exposed to other software

IRT

Institut de Recherche Technologique Saint Exupéry

jacobian

The matrix of first order partial derivatives of outputs with respect to inputs

job scheduler

A program used to distribute tasks and allocate resources for tasks on HPC clusters

JSON

JavaScript Object Notation

JSON schema

A JSON description of JSON data, similar to XML schemas

KKT

Karun Kuhn Tucker

LHS

Latin Hypercube Sampling

library wrapper

A code that translates the existing API of a program, or a library, into a compatible one

local design variables

In an MDO scenario, a subset of the design variables that are inputs of only one discipline

LOO

Leave One Out

LP

Linear Programming

LSF

COTS jobs scheduler for HPC clusters

MDA

Multi-Disciplinary Analysis

MDF

Multi-Disciplinary Feasible

MDO

Multi-Disciplinary Design Optimization

MDO architecture

the software architecture that enables the programming and resolution of MDO design problems

MDO formulation

The mathematical strategy used to define the optimization problem(s) to be solved

MDO formulations designer

A user class of GEMSEO that creates, implements, tests or maintains MDO formulations

MDO formulations engine

A program that enables the implementation of MDO formulations

MDO integrator

A user class that uses disciplines and MDO formulations to create, test and maintain MDO scenarios

MDO platform

A set of programs integrated in a common framework, enabling the resolution of multidisciplinary engineering problems using numerical simulation and optimization

MDO user

A user class of the MDO platform that executes an MDO scenario to produce results

MDOIS

MDO of Independent Subspaces

MMA

Method of Moving Asymptotes

Model Center

COTS workflow engine developed by Phoenix Integration

modular architecture

A software architecture based on separated components, with a relative independence between them

Monolithic

A type of MDO formulation which formulates a single optimization problem

Newton method

A root finding method that uses successive linear approximations of the function of interest

NLopt

Non-Linear Optimization package, an open source library of optimization algorithms from MIT, http://ab-initio.mit.edu/nlopt

Object Oriented Programming

A programming paradigm based on objects, which are data structures as well as a structure for methods

objective function

The function to be minimized or maximized in an optimization problem

OpenDACE

Open Design and Analysis of Computer Experiments: a program developed by Airbus to standardize interfaces of optimization algorithms, DOE methods and surrogate models as well as the related problems to be solved

operating condition

An input parameter of a simulation program that defines a physical parameter in which the system operates. Typically the speed of a vehicle, or its altitude

optimization algorithm

An algorithm capable of solving optimization problems

optimization history

The database of values of the objective function, constraints and design variables obtained during an optimization

optimization problem

A mathematical problem consisting in finding a set of variables which minimizes or maximizes a mathematical function (possibly a set-valued map), while satisfying constraints on these variables or on artrary functions

optimum

Solution of an optimization problem: the design variables values at the minimum of the function, subject to the constraints

PIDO

Process Integration and Design Optimization

pip

The PyPA recommended tool for installing Python packages. https://pypi.python.org/pypi/pip

process

A series of executions and data exchanges, ie the workflow and the data flow.

processes data

The disciplines inputs and generated outputs during the execution of a process

RBF

Radial Basis Function

root finding

A family of numerical resolution methods that solves multivariate problems of the type R(x)=0

run time

In computer science, run time, runtime or execution time is the time during which a program is running (executing), in contrast to other program life cycle phases such as compile time, link time and load time

SAND

Simultaneous Analysis and Design

SBO

Surrogate Based Optimization

scenario

The translation of a design problem into an executable. When executed, a scenario generates a process

Scilab

An open-source alternative to Matlab

SciPy

Open Source Library of Scientific Tools, containing a library of optimization algorithms, https://www.scipy.org

Secure Shell

(SSH) an encrypted network protocol for accessing remote computers

sequence diagram

A UML diagram that shows objects execution sequence, such as the function calls and their arguments, and the execution order

serialization

Process of writing objects or data structures to disk or more generally, to formats that can be stored

shared design variables

The design variables that are shared by more than one discipline

simulation software

A program that simulates a part of the physics of a system, or contributes to the overall simulation of the system, such as a mesher

SLM

Simulation Life Cycle Management

SLSQP

Sequential Least Squares Quadratic Programming

SOM

Self Organizing Map

SRS

Software Requirements Specification

SSBJ

SuperSonic Business Jet

surrogate model

A mathematical model of another model. Typically used to substitute an expensive simulation-based model by an approximation whose cost of evaluation is lower, at the price of an initial sampling of the original model

SWIG

Simplified Wrapper and Interface Generator

system design variables

The design variables that are shared by more than one discipline, at the system level optimization problem in a bi level MDO scenario

trade-off

A study that aims at comparing different options in terms of design parameters, and analyse their impacts

UML

Unified Modeling Language

weakly coupled problem

A multidisciplinary problem where the coupling variables can be computed by a single execution chain

work flow

The execution sequence of the disciplines in a process

workflow engine

A program used to design, run and analyze processes

workflow-driven

Characteristic of a workflow engine, for which processes are described through the work flow, and in which the data flow is deduced accordingly

WORMS

Workflow Management System

wrapper

Here discipline wrappers. Standardized interface defining inputs, outputs and execution of a given simulation software.

XDSM

eXtended Design Structure Matrix

XML

Extensible Markup Language