Charlie Vanaret

Zuse-Institut Berlin | ZIB

This paper describes the GEMS software developed as part of the IRT Saint Exupéry MDA-MDO project for supporting Multidisciplinary Design Optimization (MDO) capabilities. GEMS is a Python library for programing MDO simulation processes, built on top of NumPy, SciPy and Matplotlib. GEMS aims at pushing forward the limits of automation in simulation...

Reliable global optimization is dedicated to finding a global minimum in the presence of rounding errors. The only approaches for achieving a numerical proof of global optimality are interval branch and bound methods that interleave branching of the search-space and pruning of the subdomains that cannot contain an optimal solution. It is of the utm...

Robust optimization (RO) has attracted much attention from the optimization community over the past decade. RO is dedicated to solving optimization problems subject to uncertainty: design constraints must be satisfied for all the values of the uncertain parameters within a given uncertainty set. Uncertainty sets may be modeled as deterministic sets...

Iterative methods for nonlinear optimization usually share common ingredients, such as strategies to compute a descent direction or mechanisms that promote global convergence. We propose a four-ingredient abstract framework to describe nonlinear optimization methods and unify their workflows, and introduce Uno, a lightweight modular C++ solver that...

Iterative methods for nonlinear optimization usually share common ingredients, such as strategies to compute a descent direction or mechanisms that promote global convergence. We propose a four-ingredient abstract framework to describe nonlinear optimization methods and unify their workflows, and introduce Uno, a lightweight modular C++ solver that...

Nonlinear optimization is ubiquitous in applications as diverse as engineering design, machine learning, control, data fitting, and economic planning. Practitioners routinely use robust and efficient optimization solvers that scale up to several thousand variables and constraints. However, the immense majority of codes have been written more than 1...

This study investigates the progress made in LP and MILP solver performance during the last two decades by comparing the solver software from the beginning of the millennium with the codes available today. On average, we found out that for solving LP/MILP, computer hardware got about 20 times faster, and the algorithms improved by a factor of about...

This study investigates the progress made in LP and MILP solver performance during the last two decades by comparing the solver software from the beginning of the millennium with the codes available today. On average, we found out that for solving LP/MILP, computer hardware got about 20 times faster, and the algorithms improved by a factor of about...

The current shift of energy systems in several European countries towards sustainable, renewable technologies comes with a massive decentralization and a concomitant increase in the scale of realistic energy models. The resulting large-scale linear programs (LPs) often contain several hundred million constraints and variables. These models can take...

Within the interdisciplinary BMWK-funded project UNSEEN, experts from High Performance Computing, mathematical optimization and energy systems analysis combine strengths to evaluate uncertainties in modeling and planning future energy systems with the aid of High Performance Computing (HPC) and neural networks. Energy System Models (ESM) are centra...

Energy systems research strongly relies on large modeling frameworks. Many of them use linear optimization approaches to calculate blueprints for ideal future energy systems, which become increasingly complex and so do the models. The state-of-the-art to compute them is the application of shared-memory computers combined with approaches to reduce t...

Iterative methods for nonlinear optimization usually share common ingredients, such as strategies for computing a descent direction or mechanisms that promote global convergence. We propose a four-ingredient terminology to describe state-of-the-art methods and we introduce a new open-source framework for unifying nonlinearly constrained optimizatio...

Iterative methods for nonlinear optimization usually share common ingredients, such as strategies to compute a descent direction or mechanisms that promote global convergence. Our new open-source framework for nonlinearly constrained optimization, UNO, offers a selection of off-the-shelf strategies that can be assembled at will. UNO thus unifies a...

Der Einsatz von Modellen zur Erstellung und Untersuchung von Szenarien ist ein wesentliches Instrument der Energiesystemanalyse. Für die Politikberatung ist die Frage nach der Verlässlichkeit von solchen Szenarien von großer Wichtigkeit, da diese mit großen Unsicherheiten behaftet sein können. Diesem Problem wird in UNSEEN begegnet: durch das Abfah...

Simulation is besides experimentation the major method for designing, analyzing and optimizing chemical processes. The ability of simulations to reflect real process behavior strongly depends on model quality. Validation and adaption of process models are usually based on available plant data. Using such a model in various simulation and optimizati...

In this article, we propose an algorithmic framework for globally solving mixed problems with continuous variables and categorical variables whose properties are available from a catalog. It supports catalogs of arbitrary size and properties of arbitrary dimension, and does not require any modeling effort from the user. Our tree search approach, si...

Model-based experimental design is attracting increasing attention in chemical process engineering. Typically, an iterative procedure is pursued: an approximate model is devised, prescribed experiments are then performed and the resulting data is exploited to refine the model. To help to reduce the cost of trial-and-error approaches, strategies for...

We introduce a filter mechanism to enforce convergence for augmented Lagrangian methods for nonlinear programming. In contrast to traditional augmented Lagrangian methods, our approach does not require the use of forcing sequences that drive the first-order error to zero. Instead, we employ a filter to drive the optimality measures to zero. Our alg...

Model-based experimental design is attracting increasing attention in chemical process engineering. Typically, an iterative procedure is pursued: an approximate model is devised, prescribed experiments are then performed and the resulting data is exploited to refine the model. To help to reduce the cost of trial-and-error approaches, strategies for...

We introduce a filter mechanism to enforce convergence for augmented Lagrangian methods for nonlinear programming. In contrast to traditional augmented Lagrangian methods, our approach does not require the use of forcing sequences that drive the first-order error to zero. Instead, we employ a filter to drive the optimality measures to zero. Our alg...

Optimal (model-based) experimental design (OED) aims to determine the interactions between input and output quantities connected by an, often complicated, mathematical model as precisely as possible from a minimum number of experiments. While statistical design techniques can often be proven to be optimal for linear models, this is no longer the ca...

The purpose of an integrated mini-plant is a proof-of-concept and an intermediate upscaling between lab and production scales. A mini-plant is realized in order to identify the best operating windows for the upscaled production process and also to identify operating risks and failures. However, due to the large number of design variables and their...

The benefit of model-based process simulation and optimization to enhance and improve process development depends on the reliability of the underlying model. Its selection and the values of model parameters must be validated with experimental results. This work demonstrates that model-based design of experiments integrated in a steady-state process...

Model-based experimental design is attracting increasing attention in chemical process engineering. Typically, an iterative procedure is pursued: an approximate model is devised, prescribed experiments are then performed and the resulting data is exploited to refine the model. To help reduce the cost of trial-and-error approaches, strategies for mo...

Iterative methods for nonlinear optimization usually share common ingredients, such as strategies to compute a descent direction or mechanisms that promote global convergence. Our new open-source framework for nonlinearly constrained optimization, Argonot, offers a selection of off-the-shelf strategies that can be assembled at will. Argonot thus im...

Multidisciplinary Design Optimization (MDO) based on high-fidelity models is challenging due to the high computational cost of evaluating the objective and constraints. To choose the best MDO architecture, a trial-and-error approach is not possible due to the high cost of the overall optimization and the complexity of each implementation. We propos...

This talks presents the hybrid reliable solver that I implemented during my PhD and the new optimality results that I obtained.

Air traffic management (ATM) is an endless source of challenging optimization problems. Before discussing applications of metaheuristics to these problems, let us describe an ATM system in a few words, so that readers who are not familiar with such systems can understand the problems being addressed in this chapter. Between the moment passengers bo...

The Countdown game is one of the oldest TV show in the world. It started broadcasting in 1972 on the French television and in 1982 on British channel 4, and it has been running since in both countries. The game, while extremely popular, never received any serious scientific attention, probably because it seems too simple at first sight. We present...

The only rigorous approaches for achieving a numerical proof of optimality in global optimization are interval-based methods that interleave branching of the search-space and pruning of the subdomains that cannot contain an optimal solution. State-of-the-art solvers generally integrate local optimization algorithms to compute a good upper bound of...

The Lennard-Jones potential is a relatively realistic model that describes pairwise interactions between spherical noble gas atoms. Determining the most stable configuration of a cluster with N atoms amounts to finding the relative positions of the atoms that minimize the global potential energy ; this potential plays a crucial role in the protein...

L’optimisation globale fiable est dédiée à la recherche d’un minimum global en présence d’erreurs d’arrondis. Les seules approches fournissant une preuve numérique d’optimalité sont des méthodes d’intervalles qui partitionnent l’espace de recherche et éliminent les sous-espaces qui ne peuvent contenir de solution optimale. Ces méthodes exhaustives,...

Nonconvex and highly multimodal optimization problems represent a challenge both for stochastic and deterministic global optimization methods. The former (metaheuristics) usually achieve satisfactory solutions but cannot guarantee global optimality, while the latter (generally based on a spatial branch and bound scheme~\cite{Smith1997Global}, an ex...

http://www.uni-wuerzburg.de/fileadmin/10030000/scan2014/bookOfAbstracts.pdf#page=163

We provide the global optimization community with new optimality proofs for six deceptive benchmark functions (five bound-constrained functions and one nonlinearly constrained problem). These highly multimodal nonlinear test problems are among the most challenging benchmark functions for global optimization solvers; some have not been solved even w...

In this contribution, we explore the application of evolutionary algorithms for information filtering. There are two crucial issues we consider in this study: (1) the generation of the user’s profile which is the central task of any information filtering or routing system; (2) self-adaptation and self-evolving of the user’s profile given the dynami...

In this paper, we propose a Growing Type-2 Fuzzy Classifier (GT2FC) for online rule learning from real-time data streams. While in batch rule learning, the training data are assumed to be drawn from a stationary distribution, in online rule learning, data can dynamically change over time becoming potentially nonstationary. To accommodate dynamic ch...

When designing a wind farm layout, we can reduce the number of variables by optimizing a pattern instead of considering the position of each turbine. In this paper we show that, by reducing the problem to only two variables defining a grid, we can gain up to $3\%$ of energy output on simple examples of wind farms dealing with many turbines (up to 1...

Le chapitre présente des applications à la gestion du trafic aérien de diverses métaheuristiques : l'optimisation des routes aériennes, de l'espace, des créneaux de décollage, de la circulation sur la plateforme aéroportuaire et la détection et résolution de conflits.

Evolutionary Algorithms (EA) usually carry out an efficient exploration of the search-space, but get often trapped in local minima and do not prove the optimality of the solution. Interval-based techniques, on the other hand, yield a numerical proof of optimality of the solution. However, they may fail to converge within a reasonable time due to th...

Evolutionary Algorithms (EA) usually carry out an efficient exploration of the search-space, but get often trapped in local minima and do not prove the optimality of the solution. Interval-based techniques, on the other hand, yield a numerical proof of optimality of the solution. However, they may fail to converge within a reasonable time due to th...

Highly nonlinear and ill-conditioned numerical optimization problems take their toll on the convergence of existing resolution methods. Stochastic methods such as Evolutionary Algorithms carry out an efficient exploration of the search-space at low cost, but get often trapped in local minima and do not prove the optimality of the solution. Determin...

Stochastic algorithms have demonstrated their abilities on nonlinear numerical optimization problems, while deterministic methods struggle to converge within a reasonable time. We present a cooperative algorithm exploiting the efficiency of Evolutionary Algorithms and the reliability of Interval Constraint Programming to achieve proof of optimality...

Interval arithmetic has been used in computer science and numerical computations for years [5]. Its main goal was to create computing en-vironments where the exact value of a computed result lies with cer-tainty within an interval, which might be paramount for some critical applications. Floating point units (FPU) only work with a fixed size for th...

Applying a benchmarking approach to conflict resolution problems is a hard task, as the analytical form of the constraints is not simple. This is especially the case when using realistic dynamics and models, considering accelerating aircraft that may follow flight paths that are not direct. Currently, there is a lack of common problems and data tha...

Applying a benchmarking approach to con-flict resolution problems is a hard task, as the analytical form of the constraints is not simple. This is especially the case when using realistic dynamics and models, considering accelerating aircraft that may follow flight paths that are not direct. Currently, there is a lack of common problems and data th...

La résolution de conflits entre trajectoires d'avions est un problème fortement combinatoire qui n'a encore jamais été résolu par des méthodes déterministes locales ou globales en gardant des hypothèses réalistes. Les seules approches centralisées permettant de résoudre efficacement des conflits impliquant plus d'une vingtaine d'avions en rechercha...

Incremental learning aims at equipping data-driven systems with self-monitoring and self-adaptation mechanisms to accommodate new data in an online setting. The resulting model underlying the system can be adjusted whenever data become available. The present paper proposes a new incremental learning algorithm, called 2G2M, to learn Growing Gaussian...

Learning of fuzzy rule-based systems in dynamic environments is typical and relevant to various real world applications. Contrary to the offline rule-based systems where the process of rule induction is performed at once, incremental learning is evolutionary. Learning takes place over long periods of time and the system is subject to refinement as...