Past Events › Colloquium

On October 24, 2016,  Prof. Dr. Georg Pflug  from the University of Vienna  will join the ALOP-Colloquium and present his recent work:Multistage Stochastic Optimization: Approximation, bounds and ambiguityAbstractMultistage stochastic programs are - in contrast to two-stage programs - difficult variational problems for which solutions can be found only in exceptional cases. For this reason, we have to approximate them by easier programs. We show how a distance between these problems can be found by a variant of a transportation distance, which…

On November 7, 2016,  Prof. Dr. Maria Dolores Ugarte  from the Universidad Publica de Navarra  will join the ALOP-Colloquium and present her recent work: Title:  Analyzing cancer mortality in small areas using Bayesian B-splines  Abstract: In recent years, models incorporating splines have been considered for smoothing risks in disease mapping.  Although these models are very fexible, they can be computationally demanding in certain cases. In this talk, several models including one, two, and three dimensional B-splines (penalized or un-penalized) are proposed…

On November 28, 2016,  Dr. Benedikt Wilbertz will join the ALOP-Colloquium and present his recent work entitled Hot Topics in Machine LearningAbstractThe history of machine learning goes back to the late 50s, when the first perceptron models were created to build machines that lean from observations.It took until the mid 80s that Yann LeCun constructed a neural network which can be considered as the first production read application of modern machine learning.Since then, the field shifted its focus more towards…

On Wednesday, November 30, 2016,  Prof. Dr. Andreas Krause from the ETH Zurich will join the ALOP-Colloquium and present his recent work entitledSubmodularität und Machine Learning At 13:45 there will be a coffee hour in E-10. At 14:00 the presentation will take place in HS9.

On Thursday, January 19, 2017,  Prof. Dr. Armin Fuegenschuh from the Helmut Schmidt Universitaet Hamburg will join the ALOP-Colloquium and present his recent work entitledDesign Issues for Magnetic Rulers We deal with the design of a magnetic ruler as an absolute positioning system. Trapezoid-shaped magnetic areas are placed side by side, so that the resulting pattern enables a unique encoding of each position up to a micrometer level. The problem of finding a longest-possible ruler from a set of elementary…

On Thursday, January 23, 2017,  Dr. Roland Hildebrand from the University of Grenoble  will join the ALOP-Colloquium and present his recent work Element-wise functions preserving positivity of matrices At 15:45 there will be a coffee hour in E10. The presentation will take place in HS9 at 16:15.

On Monday, February 6, 2017,  Dr. Bart Vandereycken from the University of Geneva will join the ALOP-Colloquium and present his recent work Manifold optimization for low rank matrix and tensor completionThe minimisation of a smooth objective function subject to (matrix) rank constraints can sometimes be very effectively solved by methods from Riemannian optimisation. This is for instance the case with the low-rank matrix and tensor completion problems However, the standard theory of Riemannian optimisation leaves some questions unanswered regarding its theoretical and practical…

On Wednesday, April 19, 2017,  Dr. Marta D'Elia from the Computer Science Research Institute at Sandia National Laboratories will join the ALOP-Colloquium and present her recent work entitled A coupling strategy for nonlocal and local models with applications to static peridynamics and classical elasticityABSTRACT: The use of nonlocal models in science and engineering applications has been steadily increasing over the past decade. The ability of nonlocal theories to accurately capture effects that are difficult or impossible to represent by local Partial Differential…

On Monday, April 24, 2017,  Prof. Florian Jarre from the Heinrich Heine Universität Düsseldorf will join the ALOP-Colloquium and present his recent work "The Max-Cut-Polytop"AbstractThe max-cut-polytope arises from a simple reformulation of the problem of finding the maximum cut in a graph with given edge weights. Unfortunately, the exact pre-presentation of the cut polytope is not tractable unless NP = Co-NP.We discuss two different representations and inner and outer nonlinear approximations of the max-cut polytope leading to the anticipation that solving…

On Monday, May 15 2017,   Prof. Andreas Frommer of Bergische Universität Wuppertal will join the ALOP-Colloquium and present his recent work entitled Matrix functions - computation and convergence for Krylov subspace based methods Abstract: f(A)b, the action of the function f of a matrix A on a vector b, is required in many applications, for example in exponential integrators, where f = exp or in lattice QCD where f(z) = z-α; α ϵ (0, 1). Typically, A is large and sparse, so…

On Monday, May 29, 2017,  Prof. François Glineur from the Université catholique de Louvain will join the ALOP-Colloquium and present his recent work entitledPerformance estimation of first-order methods for composite convex optimizationComposite convex optimization consists in the minimization of a convex objective function equal to the sum of several convex functions with different properties, for example a smooth term and a nonsmooth term. We consider a large class of first-order methods designed to solve such composite problems, which rely on specific…

On Friday, June 9 2017,  Prof. Lorenz T. (Larry) Biegler,  will join the ALOP-Colloquium and present his recent work. Lorenz T. (Larry) Biegler is currently the Head and Bayer University Professor of Chemical Engineering at Carnegie Mellon University. His research interests lie in computer aided process engineering (CAPE) and include flowsheet optimization, optimization of systems of differential and algebraic equations, reactor network synthesis and algorithms for constrained, nonlinear process control. Projects in these areas include analysis and development of nonlinear programming…

On Monday, June 12, 2017,  Prof. Gabriel Wittum of King Abdullah University of Science and Technology  will join the ALOP-Colloquium and present his recent work entitledSimulation of Neuronal Signal ProcessingThe crucial feature of neuronal ensembles is their high complexity and variability. This makes modelling and computation very difficult, in particular for detailed models based on first principles. The problem starts with modelling geometry, which has to extract the essential features from those highly complex and variable phenotypes and at the…

On Monday, June 26 2017,  Dr. Giorgio Ferrari with Bielefeld University will join the ALOP-Colloquium and present his recent work entitled On the Optimal Management of Public Debt: a Singular Stochastic Control Problem Consider the problem of a government that wants to reduce the debt-to-GDP (gross domestic product) ratio of a country. The government aims at choosing a debt reduction policy which minimises the total expected cost of having debt, plus the total expected cost of interventions on the debt ratio.…

On Friday, September 8, 2017, Prof. Tim Kelley of North Carolina State University  will join the ALOP-Colloquium and present his recent work entitled   Anderson Acceleration   Abstract:   We will discuss recent results on the local convergence of Anderson Acceleration. This is an algorithm for accelerating the convergence of fixed point or Picard iteration. In this talk we will discuss the first convergence results for the method, illustrate the results with an application to radiative transport, discuss issues with…

On Monday, October 23, 2017, at 16:00 Prof. MK (Mapundi) Banda of the University of Pretoria will join the ALOP-Colloquium and present his recent work entitledOptimal Control, Stabilization and Parameter Identification for evolution modelsEvolution models governed by the time-dependent partial differential equations will be considered. In this talk, we will present some ideas on optimal control, boundary stabilization and parameter identification for a selection of such models.  For instance, boundary information will be used to stabilise or to identify parameters of the…

On Monday, November 6, 2017,  Prof. Dr. Boris Khoromskij from the Max-Planck-Institute für Mathematik in den Naturwissenschaften will join the ALOP-Colloquium and present his recent work entitledTensor numerical methods for multi-dimensional PDE's and in data modelingon Monday, November 6, 2017 at 17:00.  The presentation will take place in HS9.

On Monday, November 20, 2017, at 16:00  Prof. John Pearson of School of Mathematics, The University of Edinburgh will join the ALOP-Colloquium and present his recent work entitledFour Perspectives on Preconditioning PDE-Constrained Optimization ProblemsAbstract:PDE-constrained optimization problems have a wide range of applications across numerical mathematics and applied science, so it is important to develop fast and feasible methods to solve such problems. We employ preconditioned iterative methods to tackle the large and sparse matrix systems that arise from their discretization,…

On Monday, December 4, 2017, at 16:00 Prof. Ekaterina Kostina of Heidelberg University will join the ALOP-Colloquium and present her recent work entitled Optimization-Based Model Validation - a Mathematical Technology of High ImpactValidated dependable mathematical models are of ever growing importance in science and engineering, and today also driven by a serious demand of the industry. Apart from providing scientific insight into complex nonlinear processes, mathematical models are fundamental for process simulation, optimization and control. In this context the new paradigmatic…

On Monday, January 22, 2018, at 16:00 Prof. Andrea Walther of Paderborn University will join the ALOP-Colloquium and present her recent work entitled:Recent progress in the optimization of piecewise smooth functionsIn a previous paper, we derived first and second order convexity conditions for a large class  of piecewise smooth functions.In that analysis based on piecewise linearization, a key assumption was the Linear Independence Kink Qualification (LIKQ), a generalization of the Linear Independence Constraint Qualification (LICQ) known from smooth nonlinear optimization.…

On Tuesday, February 6, 2018, at 12:00  Prof. Fredi Tröltzsch of TU Berlin will present his recent work entitled Optimal and Feedback Control of some Reaction-Diffusion EquationsThe optimal control of some systems of nonlinear reaction-diffusion equations is considered including several important equations of mathematical physics. In particular, equations are covered that develop traveling wave fronts, spiral waves, scroll rings, or propagating spot solutions. Well-posedness of the system, differentiability of the control-to-state mapping, and optimality conditions of first and second order are…

On Monday, April 16, 2018, at 16:00 Prof. Michael Günther of Bergische Universität Wuppertal will join the ALOP-Colloquium and present his recent work entitled Effiziente numerische Simulation multiphysikalischer Systeme Multiphysikalische Systeme werden oft durch Modelle beschrieben, die DAE-Netzwerkmodelle mit partiellen Differentialgleichungen über geeignete Randbedingungen und Quellterme koppeln. Nach Semidiskretisierung im Ort hat man es dann mit großen, gekoppelten DAE-Sytemen zu tun. In diesem Vortrag werden verschiedene Ansätze zu deren effizienter numerischer Simulation besprochen: Multirate-Verfahren und dynamsche Iteration.   The presentation…

On Monday, May 14, 2018, at 16:00,  Prof. Partha Lahiri of the University of Maryland College Park will join the ALOP-Colloquium and present his recent work entitled "Big data in official statistics." Abstract:  The demand for various socio-economic, transportation, and health statistics for small geographical areas is steadily increasing at a time when survey agencies are desperately looking for ways to reduce costs to meet fixed budgetary requirements.  In the current survey environment, the application of standard sample survey methods…

On Monday, October 29, 2018 at 16:00 c.t. Prof. Dr. Ir. Arie M.C. A. Koster  of RWTH Aachen will present his recent work entitled Robust Combinatorial Optimization - New Problems and Solutions Abstract: The growing need to deal with uncertainties in optimization processes has lead to many new research directions. In robust optimization, an uncertainty set has to be defined containing all scenarios to be considered. Availability of historical data often suggests the use of a discrete or polyhedral uncertainty…

On Monday, November 26, 2018 at 16:00 c.t.  Caroline Geiersbach of Universität Wien will present her recent work entitled   Stochastic Approximation for Shape Optimization     The presentation will take place in HS 9. Please join us for coffee at 15:45 in E10.

On Monday, December 3, 2018 at 16:00 c.t. Prof. Dr. Frank Vallentin of Universität Köln will present his recent work entitled "Chromatic numbers of geometric graphs" Abstract: A classical problem in discrete geometry (due to Hadwiger and Nelson) is to find the minimal number of colors one needs to color all points in the Euclidean plane so that no two points which are distance 1 apart receive the same color. Similar geometric coloring problems can be posed in the context…

On Monday, January 28, 2019 at 16:00 c.t.  Prof. Dr. Martin Gugat of FAU Erlangen will present his recent work entitled   On the limits of boundary stabilizability for networks of strings.   Abstract: An example by Bastin and Coron illustrates that theboundary stabilization of 1-d hyperbolic systemswith certain source terms is only possible if the lengthof the space interval is sufficiently small. We show that simlar phenomena also occur fornetworks of vibrating strings that are governedby the wave equation with…

On Monday, April 29 2019 at 16:00 c.t. Prof. Florian Jarre  of  Universität Düsseldorf will present his recent work. Stochastic gradient descent und Block coordinate descent -- eine Einführung in die Theorie und die praktische Anwendung   Der Vortrag befasst sich mit extrem hoch-dimensionalen speziell strukturierten Minimierungsproblemen, für die selbst die Auswertung eines einzelnen Funktionswertes oft so viel Rechenzeit beansprucht, dass neue iterative Verfahren zum Einsatz kommen, bei denen die einzelnen Iterationen billiger sind als eine Funktionsauswertung. Natürlich kann man…

On Monday, May 6 2019 at 16:00 c.t. Prof. Dr. Christian Meyer of TU Dortmund will present his recent work entitled Optimal control of a non-smooth semi-linear elliptic equation   The presentation will take place in HS 9.   Please join us for coffee at 15:45 in E10.

On Monday, June 17 2019 at 16:00 c.t. Prof. Dipl.-Ing. Dr. Barbara Kaltenbacher of AAU Klagenfurt will present her recent work entitled Optimization based formulation and regularization of inverse problems Abstract:   The conventional way of formulating inverse problems such as identification of a (possibly infinite dimensional) parameter, is via some forward operator, which is the concatenation of the observation operator with the parameter-to-state-map for the underlying model. Recently, all-at-once formulations have been considered as an alternative to this reduced…

On Monday, July 1 2019 at 16:00 c.t.  Dr. Christopher Hojny, TU Darmstadt  will present his recent work entitled Symmetry Handling in Binary Programs - Combining Symretopes and Orbital Fixing   Abstract: Branch-and-bound is an established method to solve binary programs with thousands of variables in adequate time. If symmetries are present, however, even small instances may be hard to solve, because symmetric parts of the branch-and-bound tree are inspected repeatedly without providing new information. A standard way to handle symmetries…

On Monday, November 4 2019 at 16:00 c.t.  Prof. Dr. Volker Kaibel, Otto-von-Guericke-Universität Magdeburg  will present his recent work entitled   Projections in Combinatorial Optimization Formulating problems via linear inequalities and projections is at the heart of the polyhedral combinatorics approach to Combinatorial Optimization. We present some examples for which this can be achieved with surprisingly few inequalities and we give an almost pictorial proof showing that for other problems exponentially many inequalities are required.   The presentation will take…

On Wednesday, December 4, 2019 at 12:00 c.t., Prof. Dr. Stefan Weltge, Assistant Professor at Technical University Munich (TUM) will present his recent work entitled Persistency of Linear Programming Relaxations for Stable Set Abstract:   The Nemhauser-Trotter theorem states that the standard linear programming (LP) relaxation for the stable set problem has a remarkable property, also known as the (weak) persistency property:  for every optimal LP solution that assigns integer values to some variables, there exists an optimal integer solution…

On Thursday, December 19 2019 at 16:00 c.t.  Dr. Martine Labbé, Université Libre de Bruxelles  will present her recent work Title and Abstract will follow.   The presentation will take place in HS 9. Please join us for coffee at 15:45 in E10.

CANCELLED DUE TO THE CURRENT TRAVEL RESTRICTIONS IN CHINA This presentation will be rescheduled at a later time. On Monday, February 10 2020 at 16:00 c.t.  Dr. Shengfeng Zhu of East China Normal University will present his recent work entitled Finite element approximations of shape gradients with applications in shape optimization  Abstract: Shape optimization has many practical applications in science and engineering. Boundary type Eulerian derivative has been widely used in shape gradient algorithms. The distributed Eulerian derivative is seldom…

On Monday, February 24 2020 at 16:00 c.t.  Dr. Marianna de Santis, Università de Roma, DIAG  will present her recent work entitled Branch-and-bound Algorithms for structured Mixed Integer Nonlinear Programming Problems Abstract: Mixed Integer Nonlinear Programming (MINLP) is the area of optimization  that addresses nonlinear problems with continuous and integer variables.  MINLP has proven to be a powerful tool for modeling and it combines challenges from both combinatorial and nonlinear optimization. In this talk, we deal with the computation of…

On Tuesday, March 3 2020 at 16:00 c.t.  Prof. Dr. Hanno Gottschalk  , Bergische Universität Wuppertal  will present his recent work. entitled False Positives and False Negatives in Semantic Segmentation: Toward the assurance of the perception with deep CNN in atonomous driving Abstract: The application of machine learning technology in safety relevant environments poses several challenges regarding their reliability. This especially applies to modern deep learning architectures, which pretty much resemble a black box. In this talk, measures of uncertainty…