Minisymposia
The mini-symposia available are listed below. To submit a contribution for one of these mini-symposia go to the submission page and, during the process, select the mini-symposium ID as one of the keywords.
Multi-fidelity, surrogate modelling and design exploration of real world problems
Esther Andrés, Ingeniería de Sistemas para la Defensa de España, Spain, eandres@isdefe.es
Emiliano Iuliano, CIRA - the Italian Aerospace Research Center, Italy, e.iuliano@cira.it
Mini-symposium ID: symp01
Keywords: Multi-fidelity, surrogate modelling, design exploration
Numerous industrial design analysis and optimization problems require running computationally intensive and complex simulation codes to be able to screen multiple configurations. Despite the availability of advanced computing capabilities, the excessive computational cost makes it unfeasible to exclusively rely on expensive simulations (also known as high-fidelity simulations) for the purpose of data production, shape optimization, design space exploration, etc. An attempt to circumvent the computational budget is the use of surrogate models. Surrogate models are 'cheap-to-evaluate' mathematical approximations that mimic the computationally expensive response or behavior (black-box) of an original system over the complete or part of the design space; once constructed, a surrogate is used in lieu of the expensive high-fidelity model in order to predict the values of the objective functions at locations that do not belong to the set of points used during the fitting process or training phase.
The use of surrogate models allows more efficient exploration and exploitation of wider design spaces, but however the curse of dimensionality is still an obstacle for large and complex industrial design problems. Indeed, training an accurate surrogate model in high-dimensional spaces would still require a tremendously large number of high-fidelity evaluations. In order to get rid of such an issue, one approach is to adopt knowledge-based strategies to prune the design space in an efficient manner, without removing any significant feature; in addition, the integration of multiple models with variable fidelity and cost levels within the surrogate training has shown to be effective in accelerating the design optimization task.
This minisymposium aims at collecting and disseminating new ideas in surrogate modeling and surrogate-based optimization, involving multi-fidelity aspects, as well as their application to analysis and design optimization of real world problems. Emphasis is laid on the development and use of fast and efficient metamodels for real-world industrial optimization problems where multimodality, high non-linearity, non-differentiability, high dimensionality and high computational cost are expected.
Recent Advances in Numerical Optimization and Optimal Control and its Applications
M. Fernanda P. Costa, Rui M. Pereira, Sofia Lopes
Department of Mathematics and Applications, University of Minho, Portugal
(mfc@math.uminho.pt, rmp@math.uminho.pt, sofialopes@math.uminho.pt)
Mini-symposium ID: symp02
Keywords: numerical optimization, optimal control
The aim of the symposium is to bring together scientists with different backgrounds and interests working in the areas of Numerical Optimization and Optimal Control and Its applications. Contributions are welcome in both theoretical and computational aspects as well as applications in Numerical Optimization and Optimal Control.
The topics include, but are not limited to the following:
. Direct and Indirect Methods for Optimal Control
. Dynamic Programming for Optimal Control
. Linear and Nonlinear Numerical Optimization Methods
. Deterministic Methods
. Stochastic Methods
. Modeling
. Applications
Single and Multiobjective Bilevel Optimization
Maria João Alves and Carlos Henggeler Antunes
University of Coimbra
(mjalves@fe.uc.pt, ch@deec.uc.pt)
Mini-symposium ID: symp03
Keywords: bilevel optimization
Bilevel optimization problems model hierarchical non-cooperative decision processes in which the upper level decision maker (the leader) and the lower level decision maker (the follower) control different sets of variables and have their own objective functions subject to interdependent constraints. The lower level problem is embedded in the constraints of the upper level problem. Decisions are made sequentially, as the leader makes his decisions first by selecting values for his variables. The follower then reacts by optimizing his objective function(s) on the feasible choices restricted by the leader's decisions. Thus, the leader needs to consider the follower's reaction to the setting of his variables since this influences feasibility and the leader's objective function(s) value(s). Sequential decision-making processes that can be modeled by bilevel optimization problems arise in many aspects of resource planning, management and policy-making, namely the design of pricing policies.
A multiobjective bilevel problem (MOBP) may have multiple objective functions at one or both levels. A special case of MOBP is the semivectorial bilevel problem (SVBP), in which there is a single objective function at the upper level and multiple objectives at the lower level. The existence of multiple objective functions at the lower level problem adds further challenges and difficulties to a bilevel problem because the leader has to deal with the uncertainty related to the follower's reaction. For each leader's decision, the follower has a set of efficient solutions. If the leader has no (or has little) knowledge about the follower's preferences, it may be very difficult for him to anticipate the follower's choice among his efficient set.
This mini-symposium aims to offer presentations using evolutionary and deterministic approaches to deal with single and multiobjective bilevel optimization models, both addressing theory and methodology as well as applications on real-world problems.
Adjoint methods for Multi-physics, including Applications
K. Giannakoglou, National Technical University ff Athens (NTUA), Greece, kgianna@central.ntua.gr
M. Meheut,
ONERA, France, Michael.Meheut@onera.fr
Mini-symposium ID: symp04
Keywords: adjoint methods
This symposium is dedicated to the development, validation and applications of multidisciplinary adjoint-based optimization methods. Until today, adjoint-based methods were mainly used in industry for pure aerodynamic design. In order to meet in a near future, the industrial objectives in terms of competitiveness and environmental impact, multi-disciplinary aspects and couplings will become key design drivers, among other, for aircraft and engine manufacturers. The aim of this symposium is to present the progresses made on multi-disciplinary adjoint based optimization ranging from academic developments up to industrial and complex applications. In this symposium, several partners of the European MADELEINE project (www.madeleine-project.eu) will present their first results but the objective is also to have other contributions from European or international universities, research centres, SMEs and industrial companies. Due to the multi-disciplinary nature of the subject, contributions from sectors other than aeronautics are welcome.
Design support tools in industrial and scientific applications
Célio Fernandes, University of Minho, Portugal, b6642@dep.uminho.pt
João Miguel Nóbrega, University of Minho, Portugal, mnobrega@dep.uminho.pt
Nicolas R. Gauger, TU Kaiserslautern, Germany, nicolas.gauger@scicomp.uni-kl.de
Mini-symposium ID: symp05
Keywords: design support tools, industria application, scientific applications
The usual design approaches employed in industry are based on experimental trial-and-error procedures. As a consequence, the design process demands a large amount of resources (time, material and equipment), and heavily relies on the experience of the human resources involved. This framework has a considerable impact on the cost and time-to-market of new products. With the purpose of increasing the production efficiency, the use of numerical codes/optimization methodologies increased significantly over the last decades, which was additionally motivated by the development of better and faster computers that allow resorting to more realistic models. Nowadays, several free and open-source numerical modelling codes are available, which, due to the absence of licensing costs, have been motivating companies to invest in acquiring the required knowledge to use them. This symposium aims to hear, from the users/developers of open-source numerical modelling codes, the gains which are being obtained on the design process, covering not only results obtained on specific applications but also recent developments, which are expected to have a positive impact on computational modelling related activities.
EMO - Evolutionary Multi-Objective Optimization
Alexandre Delbem, University of São Paulo, São Carlos, Brazil, acbd@icmc.usp.br
Frederico Gadelha Guimaraes, Federal University of Minas Gerais, Belo Horizonte, Brazil, fredericoguimaraes@ufmg.br
Renê Pinto,
Department of Polymer Engineering, University of Minho, Guimarães, Portugal, rene@renesp.com.br
Mini-symposium ID: symp06
Keywords: EMO
In many real-world applications, several objective functions have to be optimized simultaneously, leading to a multi-objective optimization problem (MOP) for which an ideal solution rarely exists. Intead, MOPs typically admit multiple compromise solutions representing different trade-offs among the objectives. Due to their applicability to a wide range of MOPs, including black-box optimization problems, evolutionary algorithms for multiobjective optimization have given rise to an important and very active research area, known as Evolutionary Multiobjective Optimization (EMO). No continuity or differentiability assumptions are required by EMO algorithms, and problem characteristics such as nonlinearity, multimodality and stochasticity can be handled as well. Furthermore, preference information provided by a decision maker can be used to deliver a finite-size approximation to the solution set in a single optimization run.
This mini-symposia is intended to bring together researchers working in this and related areas to discuss all aspects related EMO, including (but not limited to):
• Many-objective optimization
• Interactive optimization
• Performance evaluation
• Expensive function evaluations
• Implementation aspects
• Theoretical foundations
• Preference articulation
• Constraint handling
• Handling of continuous, combinatorial or mixed-integer problems
• Stopping criteria
• Hybridization
• Test functions and benchmarking
• Algorithm selection and configuration
• Visualization
• Uncertainty handling
• Large-scale optimization
• Parallel models
• Real-world applications
Optimization under Uncertainty
Domenico Quagliarella, Centro Italiano Ricerche Aerospaziali (CIRA), Italy, d.quagliarella@cira.it
Massimiliano Vasile, University of Strathclyde, UK, massimiliano.vasile@strath.ac.uk
Nicolas R. Gauger, TU Kaiserslautern, Germany, nicolas.gauger@scicomp.uni-kl.de
Mini-symposium ID: symp07
Keywords: Uncertainty quantification, Optimization under uncertainty, Robust design, Reliability based design methods.
Robust or reliability-based optimization is acquiring an increasingly important role in the engineering and industrial fields since it allows us to tackle problems that would otherwise be out of the reach of classical optimization techniques. Real world design optimization problems often require that the solution meets stringent requirements of robustness and reliability when a significant deviation from the nominal project conditions is present in the actual operating conditions. In these cases, it is necessary to characterize the uncertainty and take it into account in the optimization process. Of different nature and equally important is the uncertainty introduced in the optimization process by the intrinsic uncertainties of the numerical prediction models employed. The continued progress and advancement of the computational capabilities of modern computer systems make it increasingly attractive the idea of introducing uncertainty quantification techniques directly into the optimization loop.
This mini-symposium aims to host research contributions, both theoretical and applicative, in the field of robust optimization or reliability based both in the presence of stochastic and epistemic sources of uncertainty. It aims to be a forum of discussion between the research community and industry on the perspectives of the rapidly growing field of optimization under uncertainty. The session will focus on both industrial applications and basic research to bring together practitioners in a field that is experiencing enormous broadening. Authors are encouraged to propose either success stories of application of optimization under uncertainty in industrial contexts or to present the latest developments of basic research in this exciting field.
Authors are invited to submit papers on any topic related to optimization under uncertainty. More specifically, they may focus one or more of the following topics:
• Industrial applications of optimization in presence of uncertainties
• Large scale problems
• Worst case scenario
• Robust design
• Reliability-based design
• Multi and many object optimization under uncertainty
• Evidence-based approaches and decision making
• Evolvable optimization under uncertainty
• Innovative/alternative approaches to optimization under uncertainty
• Methods to speed-up computationally expensive problems
Numerical simulation as a tool in product development for the industry
Joana Malheiro, Filipa Carneiro, Carlos Ribeiro
PIEP - Innovation in Polymer Engineering, University of Minho, Campus de Azurém, 4800-058 Guimarães, Portugal
(joana.malheiro@piep.pt, f.carneiro@piep.pt, cribeiro@piep.pt)
Mini-symposium ID: symp08
Keywords: product design; product manufacturing; numerical simulation
Nowadays, the use of numerical analysis by the industry has become important and necessary to develop new products and optimize manufacturing processes. Given the growing importance of numerical analysis as a support tool for the industry, several commercial software have emerged, which are able to respond to the specific needs of each area.
The aim of this symposium is to present different applications of simulation software in real cases of existing products. The product development process will be described, pointing out the benefits and limitations of the simulation, difficulties in the subsequent manufacture of the product and possible deviations in its performance. Among others, the topics include:
• Injection
• Extrusion
• Composite manufacturing (VARI, RTM, Thermoforming, Filament winding, etc...)
• Structural analysis
• Computational Fluid Dynamics
Design of polymer processing equipment: numerical simulation and optimization
Hubert Debski, Technical University of Lublin, Poland, h.debski@pollub.pl
Miroslaw Ferdynus, Technical University of Lublin, Poland, m.ferdynus@pollub.pl
Ivan Gajdos, Technical university of Kosice, Slovak, ivan.gajdos@tuke.sk
Mini-symposium ID: symp09
Keywords: design, polymer processing, numerical simulation, optimization
The design of polymer processing equipment is a complex task involving simultaneously numerical simulation and optimization procedures. Also, the polymer properties play an important role in this process.
The design of polymer processing equipment is a complex task involving simultaneously numerical simulation and optimization procedures. Also, the polymer properties play an important role in this process. This is a multidisciplinary field involving, not only the numerical analysis of the mechanical behavior of the equipment, but, also, the analysis of the polymer flow inside the equipment and subsequent processes ongoing during and after the product from polymer is made.
For this mini-symposium, we invite researchers involved in simulation calculations and optimization of processing, e.g. extrusion, injection molding, rotomoulding, thermoforming, as well as simulation of plastics flow in the plasticizing system, in injection channels and molding cavities or strength and thermal calculations etc...
This symposium aims to bring together researchers developing their work in this field.
Particle based simulation
Mikio Yamanoi, Prometech, Japan, yamanoi@prometech.co.jp
Kazuya Shibata, The University of Tokyo, Japan, shibata.kazuya@sys.t.u-tokyo.ac.jp
Mini-symposium ID: symp10
Keywords: particle based simulation, meshless, SPH, MPS, DEM
Considering CAE used for designing products, it needs to be easy-to-use for designers as well as to have features that has enough performance for the work. Conventional CAE tools use meshes, and the prediction accuracy depends on mesh quality, which makes it difficult for product designers to use easily and intuitively. Particle based technology is considered to be possible solution to the problem. Several particle simulation techniques have been proposed already, and Smoothed Particle Hydrodynamics (SPH) method, Moving Particle Semi-implicit (MPS) method and discrete Element method (DEM) are known as representatives. Nowadays, general-purpose simulation software based on the particle method has become commercially available, and further development of simulation technology and application know-how for industrial use are required. The aim of the mini-symposium is to share possibility of industrial use via not only fundamental research but also applied research.
Game Theory and Optimization: From Theory to Applications
Jacques Periaux, CIMNE/UPC, Barcelona, Spain, jperiaux@gmail.com
L. Mallozzi, University of Naples Federico II, Italy, lina.mallozzi@unina.it
C. De Nicola, University of Naples Federico II, Italy, denicola@unina.it
David Greiner, IUSIANI, Universidad de Las Palmas de Gran Canaria, Spain, david.greiner@ulpgc.es
Mini-symposium ID: symp11
The main objective of this Mini Symposium is to bring together researchers and technologists and to generate interest in delivering lectures on new approaches, in the field of game strategies and evolutionary computation for solving societal and industrial optimization problems.
The presentations will address the efficiency of game strategies like Pareto, Nash and Stackelberg games coupled to Evolutionary Algorithms (EAs) / Metaheuristics. Some lectures will cover advanced theoretical contributions in EAs and Metaheuristics, as well as their use for solving single or multi disciplinary design optimization problems in Applied Sciences and Engineering.
Authors are invited to submit papers on one or more of the following topics:
- Nash EAs
- Pareto based EAs
- Stackelberg EAs / Multilevel optimization
- Game Theory based EAs
- Hybridized optimizers
- Non-cooperative games / Cooperative solutions / Efficiency and Pareto optimality
- Science and Engineering applications and the above topics
References:
[1] J. Periaux, F. Gonzalez, D.S.C. Lee, "Evolutionary Optimization and Game Strategies for Advanced Multi-Disciplinary Design: Applications to Aeronautics", Intelligent Systems, Control and Automation: Science and Engineering, Vol. 75, Springer, 2015.
[2] Lina Mallozzi, "An application of optimization theory to the study of equilibria for games: a survey", Central European Journal of Operations Research, Volume 21 (3), pp. 523-539, 2013.
[3] D Greiner, J Periaux, JM Emperador, B Galvan, G Winter, "Game Theory Based Evolutionary Algorithms: A Review with Nash Applications in Structural Engineering Optimization Problems", Archives of Computational Methods in Engineering 24 (4), pp. 703-750, 2017.