kth.sePublications
Change search
Refine search result
1234567 1 - 50 of 1234
CiteExportLink to result list
Permanent link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Rows per page
  • 5
  • 10
  • 20
  • 50
  • 100
  • 250
Sort
  • Standard (Relevance)
  • Author A-Ö
  • Author Ö-A
  • Title A-Ö
  • Title Ö-A
  • Publication type A-Ö
  • Publication type Ö-A
  • Issued (Oldest first)
  • Issued (Newest first)
  • Created (Oldest first)
  • Created (Newest first)
  • Last updated (Oldest first)
  • Last updated (Newest first)
  • Disputation date (earliest first)
  • Disputation date (latest first)
  • Standard (Relevance)
  • Author A-Ö
  • Author Ö-A
  • Title A-Ö
  • Title Ö-A
  • Publication type A-Ö
  • Publication type Ö-A
  • Issued (Oldest first)
  • Issued (Newest first)
  • Created (Oldest first)
  • Created (Newest first)
  • Last updated (Oldest first)
  • Last updated (Newest first)
  • Disputation date (earliest first)
  • Disputation date (latest first)
Select
The maximal number of hits you can export is 250. When you want to export more records please use the Create feeds function.
  • 1.
    Abdullah Al Ahad, Muhammed
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Non-linearstates in parallel Blasius boundary layer2014Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    There is large theoretical, experimental and numerical interest in studying boundary layers, which develop around any body moving through a fluid. The simplest of these boundary layers lead to the theoretical abstraction of a so-called Blasius boundary layer, which can be derived under the assumption of a flat plate and zero external pressure gradient. The Blasius solution is characterised by a slow growth of the boundary layer in the streamwise direction. For practical purposes, in particular related to studying transition scenarios, non-linear finite-amplitude states (exact coherent states, edge states), but also for turbulence, a major simplification of the problem could be attained by removing this slow streamwise growth, and instead consider a parallel boundary layer. Parallel boundary layers are found in reality, e.g. when applying suction (asymptotic suction boundary layer) or rotation (Ekman boundary layer), but not in the Blasius case. As this is only a model which is not an exact solution to the Navier-Stokes (or boundary-layer) equations, some modifications have to be introduced into the governing equations in order for such an approach to be feasible. Spalart and Yang introduced a modification term to the governing Navier-Stokes equations in 1987. In this thesis work, we adapted the amplitude of the modification term introduced by Spalart and Yang to identify the nonlinear states in the parallel Blasius boundary layer. A final application of this modification was in determining the so-called edge states for boundary layers, previously found in the asymptotic suction boundary layer

    Download full text (pdf)
    fulltext
  • 2.
    Abedin, Arian
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Ligai, Wolmir
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Automatingand optimizing pile group design using a Genetic Algorithm2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In bridge design, a set of piles is referred to as a pile group. The design process of pile groups employed by many firms is currently manual, time consuming, and produces pile groups that are not robust against placement errors.

    This thesis applies the metaheuristic method Genetic Algorithm to automate and improve the design of pile groups for bridge column foundations. A software is developed and improved by implementing modifications to the Genetic Algorithm. The algorithm is evaluated by the pile groups it produces, using the Monte Carlo method to simulate errors for the purpose of testing the robustness. The results are compared with designs provided by the consulting firm Tyrens AB.

    The software is terminated manually, and generally takes less than half an hour to produce acceptable pile groups. The developed Genetic Algorithm Software produces pile groups that are more robust than the manually designed pile groups to which they are compared, using the Monte Carlo method. However, due to the visually disorganized designs, the pile groups produced by the algorithm may be di cult to get approved by Trafikverket. The software might require further modifications addressing this problem before it can be of practical use.

    Download full text (pdf)
    fulltext
  • 3. Achdou, Yves
    et al.
    Dao, Manh Khang
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Ley, Olivier
    Tchou, Nicoletta
    Finite horizon mean field games on networks2020In: Calculus of Variations and Partial Differential Equations, ISSN 0944-2669, E-ISSN 1432-0835, Vol. 59, no 5, article id 157Article in journal (Refereed)
    Abstract [en]

    We consider finite horizon stochastic mean field games in which the state space is a network. They are described by a system coupling a backward in time Hamilton-Jacobi-Bellman equation and a forward in time Fokker-Planck equation. The value functionuis continuous and satisfies general Kirchhoff conditions at the vertices. The densitymof the distribution of states satisfies dual transmission conditions: in particular,mis generally discontinuous across the vertices, and the values ofmon each side of the vertices satisfy some compatibility conditions. The stress is put on the case when the Hamiltonian is Lipschitz continuous. Existence, uniqueness and regularity results are proven.

  • 4. Adams, Henry
    et al.
    Tausz, Andrew
    Vejdemo-Johansson, Mikael
    KTH, School of Computer Science and Communication (CSC), Computer Vision and Active Perception, CVAP. Institut Jozef Stefan, Slovenia .
    javaPlex: A Research Software Package for Persistent (Co) Homology2014Conference paper (Refereed)
    Abstract [en]

    The computation of persistent homology has proven a fundamental component of the nascent field of topological data analysis and computational topology. We describe a new software package for topological computation, with design focus on needs of the research community. This tool, replacing previous jPlex and Plex, enables researchers to access state of the art algorithms for persistent homology, cohomology, hom complexes, filtered simplicial complexes, filtered cell complexes, witness complex constructions, and many more essential components of computational topology. We describe, herewithin, the design goals we have chosen, as well as the resulting software package, and some of its more novel capabilities.

  • 5.
    Adler, Jonas
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.).
    Data-driven Methods in Inverse Problems2019Doctoral thesis, comprehensive summary (Other academic)
    Abstract [en]

    In this thesis on data-driven methods in inverse problems we introduce several new methods to solve inverse problems using recent advancements in machine learning and specifically deep learning. The main goal has been to develop practically applicable methods, scalable to medical applications and with the ability to handle all the complexities associated with them.

    In total, the thesis contains six papers. Some of them are focused on more theoretical questions such as characterizing the optimal solutions of reconstruction schemes or extending current methods to new domains, while others have focused on practical applicability. A significant portion of the papers also aim to bringing knowledge from the machine learning community into the imaging community, with considerable effort spent on translating many of the concepts. The papers have been published in a range of venues: machine learning, medical imaging and inverse problems.

    The first two papers contribute to a class of methods now called learned iterative reconstruction where we introduce two ways of combining classical model driven reconstruction methods with deep neural networks. The next two papers look forward, aiming to address the question of "what do we want?" by proposing two very different but novel loss functions for training neural networks in inverse problems. The final papers dwelve into the statistical side, one gives a generalization of a class of deep generative models to Banach spaces while the next introduces two ways in which such methods can be used to perform Bayesian inversion at scale.

    Download full text (pdf)
    fulltext
  • 6.
    Adler, Jonas
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    GPU Monte Carlo scatter calculations for Cone Beam Computed Tomography2014Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    A GPU Monte Carlo code for x-ray photon transport has been implemented and extensively tested. The code is intended for scatter compensation of cone beam computed tomography images.

    The code was tested to agree with other well known codes within 5% for a set of simple scenarios. The scatter compensation was also tested using an artificial head phantom. The errors in the reconstructed Hounsfield values were reduced by approximately 70%.

    Several variance reduction methods have been tested, although most were found infeasible on GPUs. The code is nonetheless fast, and can simulate approximately 3 ·109 photons per minute on a NVIDIA Quadro 4000 graphics card. With the use of appropriate filtering methods, the code can be used to calculate patient specific scatter distributions for a full CBCT scan in approximately one minute, allowing scatter reduction in clinical applications.

    Download full text (pdf)
    fulltext
  • 7.
    Adler, Jonas
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.). Elekta.
    Lunz, Sebastian
    Centre for Mathematical Sciences, University of Cambridge, Cambridge CB3 0WA, United Kingdom.
    Verdier, Olivier
    Department of Mathematics, KTH Royal Institute of Technology, 100 44 Stockholm, Sweden ; Department of Computing, Mathematics and Physics, Western Norway University of Applied Sciences, Bergen, Norway.
    Schönlieb, Carola-Bibiane
    Centre for Mathematical Sciences, University of Cambridge, Cambridge CB3 0WA, United Kingdom.
    Öktem, Ozan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.).
    Task adapted reconstruction for inverse problemsManuscript (preprint) (Other academic)
    Abstract [en]

    The paper considers the problem of performing a task defined on a model parameter that is only observed indirectly through noisy data in an ill-posed inverse problem. A key aspect is to formalize the steps of reconstruction and task as appropriate estimators (non-randomized decision rules) in statistical estimation problems. The implementation makes use of (deep) neural networks to provide a differentiable parametrization of the family of estimators for both steps. These networks are combined and jointly trained against suitable supervised training data in order to minimize a joint differentiable loss function, resulting in an end-to-end task adapted reconstruction method. The suggested framework is generic, yet adaptable, with a plug-and-play structure for adjusting both the inverse problem and the task at hand. More precisely, the data model (forward operator and statistical model of the noise) associated with the inverse problem is exchangeable, e.g., by using neural network architecture given by a learned iterative method. Furthermore, any task that is encodable as a trainable neural network can be used. The approach is demonstrated on joint tomographic image reconstruction, classification and joint tomographic image reconstruction segmentation.

    Download full text (pdf)
    fulltext
  • 8.
    Adler, Jonas
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.). Elekta, Box 7593, 103 93 Stockholm, Sweden.
    Ringh, Axel
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Öktem, Ozan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.).
    Karlsson, Johan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Learning to solve inverse problems using Wasserstein lossManuscript (preprint) (Other academic)
    Abstract [en]

    We propose using the Wasserstein loss for training in inverse problems. In particular, we consider a learned primal-dual reconstruction scheme for ill-posed inverse problems using the Wasserstein distance as loss function in the learning. This is motivated by miss-alignments in training data, which when using standard mean squared error loss could severely degrade reconstruction quality. We prove that training with the Wasserstein loss gives a reconstruction operator that correctly compensates for miss-alignments in certain cases, whereas training with the mean squared error gives a smeared reconstruction. Moreover, we demonstrate these effects by training a reconstruction algorithm using both mean squared error and optimal transport loss for a problem in computerized tomography.

  • 9.
    Adler, Jonas
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.). Elekta.
    Öktem, Ozan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.).
    Deep Bayesian InversionManuscript (preprint) (Other academic)
    Abstract [en]

    Characterizing statistical properties of solutions of inverse problems is essential for decision making. Bayesian inversion offers a tractable framework for this purpose, but current approaches are computationally unfeasible for most realistic imaging applications in the clinic. We introduce two novel deep learning based methods for solving large-scale inverse problems using Bayesian inversion: a sampling based method using a WGAN with a novel mini-discriminator and a direct approach that trains a neural network using a novel loss function. The performance of both methods is demonstrated on image reconstruction in ultra low dose 3D helical CT. We compute the posterior mean and standard deviation of the 3D images followed by a hypothesis test to assess whether a "dark spot" in the liver of a cancer stricken patient is present. Both methods are computationally efficient and our evaluation shows very promising performance that clearly supports the claim that Bayesian inversion is usable for 3D imaging in time critical applications.

    Download full text (pdf)
    fulltext
  • 10.
    af Klinteberg, Ludvig
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Computational methods for microfluidics2013Licentiate thesis, comprehensive summary (Other academic)
    Abstract [en]

    This thesis is concerned with computational methods for fluid flows on the microscale, also known as microfluidics. This is motivated by current research in biological physics and miniaturization technology, where there is a need to understand complex flows involving microscale structures. Numerical simulations are an important tool for doing this.

    The first paper of the thesis presents a numerical method for simulating multiphase flows involving insoluble surfactants and moving contact lines. The method is based on an explicit interface tracking method, wherein the interface between two fluids is decomposed into segments, which are represented locally on an Eulerian grid. The framework of this method provides a natural setting for solving the advection-diffusion equation governing the surfactant concentration on the interface. Open interfaces and moving contact lines are also incorporated into the method in a natural way, though we show that care must be taken when regularizing interface forces to the grid near the boundary of the computational domain.

    In the second paper we present a boundary integral formulation for sedimenting particles in periodic Stokes flow, using the completed double layer boundary integral formulation. The long-range nature of the particle-particle interactions lead to the formulation containing sums which are not absolutely convergent if computed directly. This is solved by applying the method of Ewald summation, which in turn is computed in a fast manner by using the FFT-based spectral Ewald method. The complexity of the resulting method is O(N log N), as the system size is scaled up with the number of discretization points N. We apply the method to systems of sedimenting spheroids, which are discretized using the Nyström method and a basic quadrature rule.

    The Ewald summation method used in the boundary integral method of the second paper requires a decomposition of the potential being summed. In the introductory chapters of the thesis we present an overview of the available methods for creating Ewald decompositions, and show how the methods and decompositions can be related to each other.

    Download full text (pdf)
    fulltext
  • 11.
    af Klinteberg, Ludvig
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Ewald summation for the rotlet singularity of Stokes flow2016Report (Other academic)
    Abstract [en]

    Ewald summation is an efficient method for computing the periodic sums that appear when considering the Green's functions of Stokes flow together with periodic boundary conditions. We show how Ewald summation, and accompanying truncation error estimates, can be easily derived for the rotlet, by considering it as a superposition of electrostatic force calculations.

    Download full text (pdf)
    fulltext
  • 12.
    af Klinteberg, Ludvig
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Fast and accurate integral equation methods with applications in microfluidics2016Doctoral thesis, comprehensive summary (Other academic)
    Abstract [en]

    This thesis is concerned with computational methods for fluid flows on the microscale, also known as microfluidics. This is motivated by current research in biological physics and miniaturization technology, where there is a need to understand complex flows involving microscale structures. Numerical simulations are an important tool for doing this.

    The first, and smaller, part of the thesis presents a numerical method for simulating multiphase flows involving insoluble surfactants and moving contact lines. The method is based on an interface decomposition resulting in local, Eulerian grid representations. This provides a natural setting for solving the PDE governing the surfactant concentration on the interface.

    The second, and larger, part of the thesis is concerned with a framework for simulating large systems of rigid particles in three-dimensional, periodic viscous flow using a boundary integral formulation. This framework can solve the underlying flow equations to high accuracy, due to the accurate nature of surface quadrature. It is also fast, due to the natural coupling between boundary integral methods and fast summation methods.

    The development of the boundary integral framework spans several different fields of numerical analysis. For fast computations of large systems, a fast Ewald summation method known as Spectral Ewald is adapted to work with the Stokes double layer potential. For accurate numerical integration, a method known as Quadrature by Expansion is developed for this same potential, and also accelerated through a scheme based on geometrical symmetries. To better understand the errors accompanying this quadrature method, an error analysis based on contour integration and calculus of residues is carried out, resulting in highly accurate error estimates.

    Download full text (pdf)
    fulltext
  • 13.
    af Klinteberg, Ludvig
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Saffar Shamshirgar, Davoud
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Tornberg, Anna-Karin
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Fast Ewald summation for free-space Stokes potentials2017In: Research in the Mathematical Sciences, ISSN 2197-9847, Vol. 4, no 1Article in journal (Refereed)
    Abstract [en]

    We present a spectrally accurate method for the rapid evaluation of free-space Stokes potentials, i.e., sums involving a large number of free space Green’s functions. We consider sums involving stokeslets, stresslets and rotlets that appear in boundary integral methods and potential methods for solving Stokes equations. The method combines the framework of the Spectral Ewald method for periodic problems (Lindbo and Tornberg in J Comput Phys 229(23):8994–9010, 2010. doi: 10.1016/j.jcp.2010.08.026 ), with a very recent approach to solving the free-space harmonic and biharmonic equations using fast Fourier transforms (FFTs) on a uniform grid (Vico et al. in J Comput Phys 323:191–203, 2016. doi: 10.1016/j.jcp.2016.07.028 ). Convolution with a truncated Gaussian function is used to place point sources on a grid. With precomputation of a scalar grid quantity that does not depend on these sources, the amount of oversampling of the grids with Gaussians can be kept at a factor of two, the minimum for aperiodic convolutions by FFTs. The resulting algorithm has a computational complexity of $$O(N \log N)$$ O ( N log N ) for problems with N sources and targets. Comparison is made with a fast multipole method to show that the performance of the new method is competitive.

  • 14.
    af Klinteberg, Ludvig
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Tornberg, Anna-Karin
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    A fast integral equation method for solid particles in viscous flow using quadrature by expansionManuscript (preprint) (Other academic)
    Abstract [en]

    Boundary integral methods are advantageous when simulating viscous flow around rigid particles, due to the reduction in number of unknowns and straightforward handling of the geometry. In this work we present a fast and accurate framework for simulating spheroids in periodic Stokes flow, which is based on the completed double layer boundary integral formulation. The framework implements a new method known as quadrature by expansion (QBX), which uses surrogate local expansions of the layer potential to evaluate it to very high accuracy both on and off the particle surfaces. This quadrature method is accelerated through a newly developed precomputation scheme. The long range interactions are computed using the spectral Ewald (SE) fast summation method, which after integration with QBX allows the resulting system to be solved in M log M time, where M is the number of particles. This framework is suitable for simulations of large particle systems, and can be used for studying e.g. porous media models.

    Download full text (pdf)
    fulltext
  • 15.
    af Klinteberg, Ludvig
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Tornberg, Anna-Karin
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Adaptive Quadrature by Expansion for Layer Potential Evaluation in Two Dimensions2018In: SIAM Journal on Scientific Computing, ISSN 1064-8275, E-ISSN 1095-7197, Vol. 40, no 3, p. A1225-A1249Article in journal (Refereed)
    Abstract [en]

    When solving partial differential equations using boundary integral equation methods, accurate evaluation of singular and nearly singular integrals in layer potentials is crucial. A recent scheme for this is quadrature by expansion (QBX), which solves the problem by locally approximating the potential using a local expansion centered at some distance from the source boundary. In this paper we introduce an extension of the QBX scheme in two dimensions (2D) denoted AQBX—adaptive quadrature by expansion—which combines QBX with an algorithm for automated selection of parameters, based on a target error tolerance. A key component in this algorithm is the ability to accurately estimate the numerical errors in the coefficients of the expansion. Combining previous results for flat panels with a procedure for taking the panel shape into account, we derive such error estimates for arbitrarily shaped boundaries in 2D that are discretized using panel-based Gauss–Legendre quadrature. Applying our scheme to numerical solutions of Dirichlet problems for the Laplace and Helmholtz equations, and also for solving these equations, we find that the scheme is able to satisfy a given target tolerance to within an order of magnitude, making it useful for practical applications. This represents a significant simplification over the original QBX algorithm, in which choosing a good set of parameters can be hard.

    Download full text (pdf)
    fulltext
  • 16.
    af Klinteberg, Ludvig
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Tornberg, Anna-Karin
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Estimation of quadrature errors in layer potential evaluation using quadrature by expansionManuscript (preprint) (Other academic)
    Abstract [en]

    In boundary integral methods it is often necessary to evaluate layer potentials on or close to the boundary, where the underlying integral is difficult to evaluate numerically. Quadrature by expansion (QBX) is a new method for dealing with such integrals, and it is based on forming a local expansion of the layer potential close to the boundary. In doing so, one introduces a new quadrature error due to nearly singular integration in the evaluation of expansion coefficients. Using a method based on contour integration and calculus of residues, the quadrature error of nearly singular integrals can be accurately estimated. This makes it possible to derive accurate estimates for the quadrature errors related to QBX, when applied to layer potentials in two and three dimensions. As examples we derive estimates for the Laplace and Helmholtz single layer potentials. These results can be used for parameter selection in practical applications.

    Download full text (pdf)
    fulltext
  • 17.
    af Klinteberg, Ludvig
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Tornberg, Anna-Karin
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Fast Ewald summation for Stokesian particle suspensions2014In: International Journal for Numerical Methods in Fluids, ISSN 0271-2091, E-ISSN 1097-0363, Vol. 76, no 10, p. 669-698Article in journal (Refereed)
    Abstract [en]

    We present a numerical method for suspensions of spheroids of arbitrary aspect ratio, which sediment under gravity. The method is based on a periodized boundary integral formulation using the Stokes double layer potential. The resulting discrete system is solved iteratively using generalized minimal residual accelerated by the spectral Ewald method, which reduces the computational complexity to O(N log N), where N is the number of points used to discretize the particle surfaces. We develop predictive error estimates, which can be used to optimize the choice of parameters in the Ewald summation. Numerical tests show that the method is well conditioned and provides good accuracy when validated against reference solutions. 

    Download full text (pdf)
    postprint
  • 18.
    Agrawal, Vishal
    et al.
    KTH, School of Engineering Sciences (SCI), Engineering Mechanics, Fluid Mechanics and Engineering Acoustics. KTH, School of Engineering Sciences (SCI), Centres, Linné Flow Center, FLOW.
    Kulachenko, Artem
    KTH, School of Engineering Sciences (SCI), Engineering Mechanics, Vehicle Engineering and Solid Mechanics, Solid Mechanics.
    Scapin, Nicolo
    KTH, School of Engineering Sciences (SCI), Engineering Mechanics, Fluid Mechanics and Engineering Acoustics. KTH, School of Engineering Sciences (SCI), Centres, Linné Flow Center, FLOW.
    Tammisola, Outi
    KTH, School of Engineering Sciences (SCI), Centres, Linné Flow Center, FLOW. KTH, School of Engineering Sciences (SCI), Engineering Mechanics, Fluid Mechanics and Engineering Acoustics.
    Brandt, Luca
    KTH, School of Engineering Sciences (SCI), Engineering Mechanics, Fluid Mechanics and Engineering Acoustics. KTH, School of Engineering Sciences (SCI), Centres, Linné Flow Center, FLOW. Department of Energy and Process Engineering, Norwegian University of Science and Technology, Trondheim, Norway.
    An efficient isogeometric/finite-difference immersed boundary method for the fluid–structure interactions of slender flexible structures2024In: Computer Methods in Applied Mechanics and Engineering, ISSN 0045-7825, E-ISSN 1879-2138, Vol. 418, article id 116495Article in journal (Refereed)
    Abstract [en]

    In this contribution, we present a robust and efficient computational framework capable of accurately capturing the dynamic motion and large deformation/deflection responses of highly-flexible rods interacting with an incompressible viscous flow. Within the partitioned approach, we adopt separate field solvers to compute the dynamics of the immersed structures and the evolution of the flow field over time, considering finite Reynolds numbers. We employ a geometrically exact, nonlinear Cosserat rod formulation in the context of the isogeometric analysis (IGA) technique to model the elastic responses of each rod in three dimensions (3D). The Navier–Stokes equations are resolved using a pressure projection method on a standard staggered Cartesian grid. The direct-forcing immersed boundary method is utilized for coupling the IGA-based structural solver with the finite-difference fluid solver. In order to fully exploit the accuracy of the IGA technique for FSI simulations, the proposed framework introduces a new procedure that decouples the resolution of the structural domain from the fluid grid. Uniformly distributed Lagrangian markers with density relative to the Eulerian grid are generated to communicate between Lagrangian and Eulerian grids consistently with IGA. We successfully validate the proposed computational framework against two- and three-dimensional FSI benchmarks involving flexible filaments undergoing large deflections/motions in an incompressible flow. We show that six times coarser structural mesh than the flow Eulerian grid delivers accurate results for classic benchmarks, leading to a major gain in computational efficiency. The simultaneous spatial and temporal convergence studies demonstrate the consistent performance of the proposed framework, showing that it conserves the order of the convergence, which is the same as that of the fluid solver.

  • 19.
    Aguilar, Xavier
    et al.
    KTH, Centres, SeRC - Swedish e-Science Research Centre. KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Computational Science and Technology (CST).
    Jordan, H.
    Heller, T.
    Hirsch, A.
    Fahringer, T.
    Laure, Erwin
    KTH, Centres, SeRC - Swedish e-Science Research Centre. KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Computational Science and Technology (CST).
    An On-Line Performance Introspection Framework for Task-Based Runtime Systems2019In: 19th International Conference on Computational Science, ICCS 2019, Springer Verlag , 2019, p. 238-252Conference paper (Refereed)
    Abstract [en]

    The expected high levels of parallelism together with the heterogeneity and complexity of new computing systems pose many challenges to current software. New programming approaches and runtime systems that can simplify the development of parallel applications are needed. Task-based runtime systems have emerged as a good solution to cope with high levels of parallelism, while providing software portability, and easing program development. However, these runtime systems require real-time information on the state of the system to properly orchestrate program execution and optimise resource utilisation. In this paper, we present a lightweight monitoring infrastructure developed within the AllScale Runtime System, a task-based runtime system for extreme scale. This monitoring component provides real-time introspection capabilities that help the runtime scheduler in its decision-making process and adaptation, while introducing minimum overhead. In addition, the monitoring component provides several post-mortem reports as well as real-time data visualisation that can be of great help in the task of performance debugging.

  • 20.
    Ahlberg, Marcus
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Fornander, Eric
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Test Case Prioritization as a Mathematical Scheduling Problem2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Software testing is an extremely important phase of product development where the objective is to detect hidden bugs. The usually high complexity of today’s products makes the testing very resource intensive since numerous test cases have to be generated in order to detect all potential faults. Therefore, improved strategies of the testing process is of high interest for many companies. One area where there exists potential for improvement is the order by which test cases are executed to detect faults as quickly as possible, which in research is known as the test case prioritization problem. In this thesis, an extension to this problem is studied where dependencies between test cases are present and the processing times of the test cases are known. As a first result of the thesis, a mathematical model of the test case prioritization problem with dependencies and known processing times as a mathematical scheduling problem is presented. Three different solution algorithms to this problem are subsequently evaluated: A Sidney decomposition algorithm, an own-designed heuristic algorithm and an algorithm based on Smith’s rule. The Sidney decomposition algorithm outper-formed the others in terms of execution time of the algorithm and objective value of the generated schedule. The evaluation was conducted by simulation with artificial test suites and via a case study in industry through a company in the railway domain.

    Download full text (pdf)
    fulltext
  • 21.
    Ahlin, Filip
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Internal model for spread risk under Solvency II2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In May 2009 the European Commission decided on new regulations regarding solvency among insurance firms, the Solvency II Directive. The directive aims to strengthen the connection between the requirement of solvency and risks for insurance firms. The directive partly consists of a market risk module, in which a credit spread risk is a sub category.

    In this thesis a model for credit spread risk is implemented. The model is an extended version of the Jarrow, Lando and Turnbull model (A Markov Model for theTerm Structure of Credit Risk Spreads, 1997) as proposed by Dubrana (A Stochastic Model for Credit Spreads under a Risk-Neutral Framework through the use of an Extended Version of the Jarrow, Lando and Turnbull Model, 2011). The implementation includes the calibration of a stochastic credit risk driver as well as a simulation of bond returns with the allowance of credit transitions and defaults.

    The modeling will be made with the requirements of the Solvency II Directive in mind. Finally, the result will be compared with the Solvency II standard formula for the spread risk sub-module.

    Download full text (pdf)
    fulltext
  • 22.
    Ahmad, Alireza
    et al.
    KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Soil and Rock Mechanics.
    Larsson, Stefan
    KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Soil and Rock Mechanics.
    Wersäll, Carl
    KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Soil and Rock Mechanics.
    Scaling granular material with polygonal particles in discrete element modeling2023In: Particuology, ISSN 1674-2001, E-ISSN 2210-4291, Vol. 75, p. 151-164Article in journal (Refereed)
    Abstract [en]

    Despite advancements in computational resources, the discrete element method (DEM) still requires considerable computational time to solve detailed problems, especially when it comes to the large-scale models. In addition to the geometry scale of the problem, the particle shape has a dramatic effect on the computational cost of DEM. Therefore, many studies have been performed with simplified spherical particles or clumps. Particle scaling is an approach to increase the particle size to reduce the number of particles in the DEM. Although several particle scaling methods have been introduced, there are still some disagreements regarding their applicability to certain aspects of problems. In this study, the effect of particle scalping on the shear behavior of granular material is explored. Real granular particles were scanned and imported as polygonal particles in the direct shear test. The effect of particle size distribution, particle angularity, and the amount of scalping were investigated. The results show that particle scalping can simulate the correct shear behavior of the model with significant improvement in computational time. Also, the accuracy of the scalping method depends on the particle angularity and particle size range.

  • 23.
    Ahmadi-Djam, Adrian
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Belfrage Nordström, Sean
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Forecasting Non-Maturing Liabilities2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    With ever increasing regulatory pressure financial institutions are required to carefully monitor their liquidity risk. This Master thesis focuses on asserting the appropriateness of time series models for forecasting deposit volumes by using data from one undisclosed financial institution. Holt-Winters, Stochastic Factor, ARIMA and ARIMAX models are considered with the latter being the one with best out-of-sample performance. The ARIMAX model is appropriate for forecasting deposit volumes on a 3 to 6 month horizon with seasonality accounted for through monthly dummy variables. Explanatory variables such as market volatility and interest rates do improve model accuracy but vastly increases complexity due to the simulations needed for forecasting.

    Download full text (pdf)
    fulltext
  • 24.
    Ahmed, Noor
    KTH, School of Industrial Engineering and Management (ITM), Learning.
    Elevers förståelse av bråktal: som ett tal som har ett eget värde på tallinjen2021Independent thesis Advanced level (professional degree), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    The aim of this study was to get insight into how the pupils in years 7 and 8 understand equivalent fractions, reducing and raising, and into how they interpret the connections between raising and multiplication and between reducing and division. The aim was also to investigate pupils' knowledge of fractional aspects with a focus on fraction as a number that has its own value on the number line. The study was based on theoretical models of how fractions can be understood, theories of learning and previous research into similar issues. The approach chosen was multi-method research, which includes quantitative methods in the collection and analysis of questionnaires and qualitative methods in the collection and analysis of interviews with pupils. Both the pupils' solutions and answers to the questions in the questionnaire and the interviews they provided gave very useful qualitative information to analyse. The analysis answered my questions, and I obtained an insight into how the pupils in years 7 and 8 understand equivalent fractions, reducing and raising as concepts and calculation methods. Through the pupils' solutions and answers, I gleaned an insight into the way in which the pupils think and see mathematics, specifically in fractions. The study indicated that pupils have sufficient knowledge of fractions as a part of a whole, while shortfalls in knowledge were identified in some pupils not mastering the concept of fractions or what numerators and denominators represent. The study also showed that some pupils have insufficient knowledge of equivalent fractions and the fact that fractions can be written in an infinite number of ways without this changing the value. In addition, they had difficulty with raising and reducing. Understanding all these concepts is necessary for pupils to operate effectively with fractions. The areas where they lacked knowledge led them to use incorrect strategies when dealing with fractions in the data. Incorrect strategies were identified pupils using their old knowledge of natural numbers and trying to adapt the answers to the new situation.

    Download full text (pdf)
    fulltext
  • 25.
    Aho, Yousef
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Persson, Johannes
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Factors Affecting the Conversion Rate in the Flight Comparison Industry: A Logistic Regression Approach2018Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Using logistic regression, we aim to construct a model to examine the factors that are most influential in affecting user behavior on the flight comparison site flygresor.se. The factors examined were number of adults, number of children, number of stops on the inbound trip, number of stops on the outbound trip, number of days between the search date and the departure date and number of search results displayed for the user. The data sample, collected during a one-week period, was taken from Flygresor and consisted of trips to or from Sweden, made within Europe, excluding Nordic countries, and made more than six days before departure. To find the variables which best explain the user behavior, variable selection methods were used along with hypothesis testing. Also, multicollinearity analysis and residual analysis were performed to evaluate the final model. The result showed that the factor number of children had no significant impact on the conversion rate, while the remaining factors had a high impact. The final model has a high predictive ability on the user's propensity to select a certain flight.

    Download full text (pdf)
    fulltext
  • 26.
    Ait Ali, Abderrahman
    et al.
    KTH, School of Architecture and the Built Environment (ABE), Transport Science, Transport Planning, Economics and Engineering.
    Lindberg, Per Olov
    KTH.
    Nilsson, Jan-Eric
    Eliasson, Jonas
    Aronsson, Martin
    Disaggregation in Bundle Methods: Application to the Train Timetabling Problem2017Conference paper (Refereed)
    Abstract [en]

    Bundle methods are often used to solve dual problems that arise from Lagrangian relaxations of large scale optimization problems. An example of such problems is the train timetabling problem. This paper focuses on solving a dual problem that arises from Lagrangian relaxation of a train timetabling optimization program. The dual problem is solved using bundle methods. We formulate and compare the performances of two different bundle methods: the aggregate method, which is a standard method, and a new, disaggregate, method which is proposed here. The two methods were tested on realistic train timetabling scenarios from the Iron Ore railway line. The numerical results show that the new disaggregate approach generally yields faster convergence than the standard aggregate approach.

    Download full text (pdf)
    disaggregation in BM
  • 27.
    Alathur Srinivasan, Prem Anand
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Deep Learning models for turbulent shear flow2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Deep neural networks trained with spatio-temporal evolution of a dynamical system may be regarded as an empirical alternative to conventional models using differential equations. In this thesis, such deep learning models are constructed for the problem of turbulent shear flow. However, as a first step, this modeling is restricted to a simplified low-dimensional representation of turbulence physics. The training datasets for the neural networks are obtained from a 9-dimensional model using Fourier modes proposed by Moehlis, Faisst, and Eckhardt [29] for sinusoidal shear flow. These modes were appropriately chosen to capture the turbulent structures in the near-wall region. The time series of the amplitudes of these modes fully describe the evolution of flow. Trained deep learning models are employed to predict these time series based on a short input seed. Two fundamentally different neural network architectures, namely multilayer perceptrons (MLP) and long short-term memory (LSTM) networks are quantitatively compared in this work. The assessment of these architectures is based on (i) the goodness of fit of their predictions to that of the 9-dimensional model, (ii) the ability of the predictions to capture the near-wall turbulence structures, and (iii) the statistical consistency of the predictions with the test data. LSTMs are observed to make predictions with an error that is around 4 orders of magnitude lower than that of the MLP. Furthermore, the flow fields constructed from the LSTM predictions are remarkably accurate in their statistical behavior. In particular, deviations of 0:45 % and 2:49 % between the true data and the LSTM predictions were obtained for the mean flow and the streamwise velocity fluctuations, respectively.

    Download full text (pdf)
    fulltext
  • 28.
    Albinski, Szymon Janusz
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    A branch-and-cut method for the Vehicle Relocation Problem in the One-Way Car-Sharing2015Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The purpose of this thesis is to develop an algorithm which solves the Vehicle Relocation Problem in the One-Way Car-Sharing (VRLPOWCS) as fast as possible. The problem describes the task of relocating the cars to areas with the largest demand. The chauffeurs who relocate the cars are transported by shuttle buses. Each car is assigned an individual relocation utility. The objective is to find shuttle tours that maximise in a given time the relocation utility while balancing the distribution of the cars. The VRLPOWCS is formulated as a mixed integer linear program. Since this problem is NP-complete we choose the branch-and-cut method to solve it. Using additional cutting planes – which exploit the structure of the VRLPOWCS – we enhance this method. Tests on real data show that this extended algorithm can solve the VRLPOWCS faster.

    Download full text (pdf)
    fulltext
  • 29.
    Alexei, Iupinov
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA.
    Implementation of the Particle Mesh Ewald method on a GPU2016Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The Particle Mesh Ewald (PME) method is used for efficient long-range electrostatic calculations in molecular dynamics (MD).

    In this project, PME is implemented for a single GPU alongside the existing CPU implementation, using the code base of an open source MD software GROMACS and NVIDIA CUDA toolkit. The performance of the PME GPU implementation is then studied.

    The motivation for the project is examining the PME algorithm’s parallelism, and its potential benefit for performance scalability of MD simulations on various hardware.

    Download full text (pdf)
    fulltext
  • 30.
    Alexis, Sara
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Combinatorial and price efficient optimization of the underlying assets in basket options2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The purpose of this thesis is to develop an optimization model that chooses the optimal and price efficient combination of underlying assets for a equally weighted basket option.

    To obtain a price efficient combination of underlying assets a function that calculates the basket option price is needed, for further use in an optimization model. The closed-form basket option pricing is a great challenge, due to the lack of a distribution describing the augmented stochastic price process. Many types of approaches to price an basket option has been made. In this thesis, an analytical approximation of the basket option price has been used, where the analytical approximation aims to develop a method to describe the augmented price process. The approximation is done by moment matching, i.e. matching the first two moments of the real distribution of the basket option with an lognormal distribution. The obtained price function is adjusted and used as the objective function in the optimization model.

    Furthermore, since the goal is to obtain en equally weighted basket option, the appropriate class of optimization models to use are binary optimization problems. This kind of optimization model is in general hard to solve - especially for increasing dimensions. Three different continuous relaxations of the binary problem has been applied in order to obtain continuous problems, that are easier to solve.

    The results shows that the purpose of this thesis is fulfilled when formulating and solving the optimization problem - both as an binary and continuous nonlinear optimization model. Moreover, the results from a Monte Carlo simulation for correlated stochastic processes shows that the moment matching technique with a lognormal distribution is a good approximation for pricing a basket option.

    Download full text (pdf)
    fulltext
  • 31.
    Alfonsetti, Elisabetta
    et al.
    KTH, School of Electrical Engineering (EES).
    Weeraddana, P. C.
    Fischione, Carlo
    KTH, School of Electrical Engineering (EES), Automatic Control.
    Min-max fair car-parking slot assignment2015In: Proceedings of the WoWMoM 2015: A World of Wireless Mobile and Multimedia Networks, IEEE conference proceedings, 2015Conference paper (Refereed)
    Abstract [en]

    Empirical studies show that cruising for car parking accounts for a non-negligible amount of the daily traffic, especially in central areas of large cities. Therefore, mechanisms for minimizing traffic from cruising directly affect the dynamics of traffic congestions. One way to minimizing cruising traffic is efficient car-parking-slot assignment. Usually, the related design problems are combinatorial and the worst-case complexity of optimal methods grows exponentially with the problem sizes. As a result, almost all existing methods for parking slot assignment are simple and greedy approaches, where each car or the user is assigned a free parking slot, which is closer to its destination. Moreover, no emphasis is placed to optimize any form of fairness among the users as the a social benefit. In this paper, the fairness as a metric for modeling the aggregate social benefit of the users is considered. An algorithm based on Lagrange duality is developed for car-parking-slot assignment. Numerical results illustrate the performance of the proposed algorithm compared to the optimal assignment and a greedy method.

  • 32.
    Al-Hassan, Yazid
    KTH, School of Computer Science and Communication (CSC).
    Simulation of Suspensions of Curved Fibers.2012Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Yazid Al-Hassan

    Simulation of Suspensions of Curved Fibers

    The thesis at hand presents a numerical method for simulations of the dynamics of slender rigid fibers immersed in an incompressible fluid. The underlying mathematical formulation is based on a slender body approximation as applied to a boundary integral equation for Stokes flow. The curvature and torsion of the fibers can be arbitrarily specified, and we consider fiber shapes ranging from moderately bent to high curvature helical shapes. Two different settings are considered; naturally buoyant fibers in shear flow and heavier fibers sedimenting due to gravity. The dynamics show a very rich behavior, with fiber trajectories that display a very different degree of regularity depending on the initial conditions and fiber shape.

  • 33.
    Allerbo, Oskar
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Jörnsten, Rebecka
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Elastic Gradient Descent, an Iterative Optimization Method Approximating the Solution Paths of the Elastic Net2023In: Journal of machine learning research, ISSN 1532-4435, E-ISSN 1533-7928, Vol. 24, no 277, p. 1-53Article in journal (Refereed)
    Abstract [en]

    The elastic net combines lasso and ridge regression to fuse the sparsity property of lasso with the grouping property of ridge regression. The connections between ridge regression and gradient descent and between lasso and forward stagewise regression have previously been shown. Similar to how the elastic net generalizes lasso and ridge regression, we introduce elastic gradient descent, a generalization of gradient descent and forward stagewise regression. We theoretically analyze elastic gradient descent and compare it to the elastic net and forward stagewise regression. Parts of the analysis are based on elastic gradient flow, a piecewise analytical construction, obtained for elastic gradient descent with infinitesimal step size. We also compare elastic gradient descent to the elastic net on real and simulated data and show that it provides similar solution paths, but is several orders of magnitude faster. Compared to forward stagewise regression, elastic gradient descent selects a model that, although still sparse, provides considerably lower prediction and estimation errors. 

  • 34.
    Allerbo, Oskar
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Solving Kernel Ridge Regression with Gradient Descent for a Non-Constant Kernel2023Manuscript (preprint) (Other academic)
    Abstract [en]

    Kernel ridge regression, KRR, is a generalization of linear ridge regression that is non-linear in the data, but linear in the parameters. The solution can be obtained either as a closed-form solution, which includes a matrix inversion, or iteratively through gradient descent. Using the iterative approach opens up for changing the kernel during training, something that is investigated in this paper. We theoretically address the effects this has on model complexity and generalization. Based on our findings, we propose an update scheme for the bandwidth of translational-invariant kernels, where we let the bandwidth decrease to zero during training, thus circumventing the need for hyper-parameter selection. We demonstrate on real and synthetic data how decreasing the bandwidth during training outperforms using a constant bandwidth, selected by cross-validation and marginal likelihood maximization. We also show theoretically and empirically that using a decreasing bandwidth, we are able to achieve both zero training error in combination with good generalization, and a double descent behavior, phenomena that do not occur for KRR with constant bandwidth but are known to appear for neural networks.

  • 35.
    Allerbo, Oskar
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Solving Kernel Ridge Regression with Gradient-Based Optimization MethodsManuscript (preprint) (Other academic)
    Abstract [en]

    Kernel ridge regression, KRR, is a generalization of linear ridge regression that is non-linear in the data, but linear in the parameters. Here, we introduce an equivalent formulation of the objective function of KRR, opening up both for using penalties other than the ridge penalty and for studying kernel ridge regression from the perspective of gradient descent. Using a continuous-time perspective, we derive a closed-form solution for solving kernel regression with gradient descent, something we refer to as kernel gradient flow, KGF, and theoretically bound the differences between KRR and KGF, where, for the latter, regularization is obtained through early stopping. We also generalize KRR by replacing the ridge penalty with the ℓ1 and ℓ∞ penalties, respectively, and use the fact that analogous to the similarities between KGF and KRR, ℓ1 regularization and forward stagewise regression (also known as coordinate descent), and ℓ∞ regularization and sign gradient descent, follow similar solution paths. We can thus alleviate the need for computationally heavy algorithms based on proximal gradient descent. We show theoretically and empirically how the ℓ1 and ℓ∞ penalties, and the corresponding gradient-based optimization algorithms, produce sparse and robust kernel regression solutions, respectively. 

  • 36.
    Allerbo, Oskar
    et al.
    University of Gothenburg and Chalmers University of Technology, Gothenburg, Sweden.
    Jörnsten, Rebecka
    University of Gothenburg and Chalmers University of Technology, Gothenburg, Sweden.
    Flexible, non-parametric modeling using regularized neural networks2022In: Computational statistics (Zeitschrift), ISSN 0943-4062, E-ISSN 1613-9658, Vol. 37, no 4, p. 2029-2047Article in journal (Refereed)
    Abstract [en]

    Non-parametric, additive models are able to capture complex data dependencies in a flexible, yet interpretable way. However, choosing the format of the additive components often requires non-trivial data exploration. Here, as an alternative, we propose PrAda-net, a one-hidden-layer neural network, trained with proximal gradient descent and adaptive lasso. PrAda-net automatically adjusts the size and architecture of the neural network to reflect the complexity and structure of the data. The compact network obtained by PrAda-net can be translated to additive model components, making it suitable for non-parametric statistical modelling with automatic model selection. We demonstrate PrAda-net on simulated data, where we compare the test error performance, variable importance and variable subset identification properties of PrAda-net to other lasso-based regularization approaches for neural networks. We also apply PrAda-net to the massive U.K. black smoke data set, to demonstrate how PrAda-net can be used to model complex and heterogeneous data with spatial and temporal components. In contrast to classical, statistical non-parametric approaches, PrAda-net requires no preliminary modeling to select the functional forms of the additive components, yet still results in an interpretable model representation. 

  • 37.
    Allerbo, Oskar
    et al.
    Chalmers university.
    Jörnsten, Rebecka
    Chalmers university.
    Non-linear, Sparse Dimensionality Reduction via Path Lasso Penalized Autoencoders2021In: Journal of machine learning research, ISSN 1532-4435, E-ISSN 1533-7928, Vol. 22, p. 1-28Article in journal (Refereed)
    Abstract [en]

    High-dimensional data sets are often analyzed and explored via the construction of a latent low-dimensional space which enables convenient visualization and efficient predictive modeling or clustering. For complex data structures, linear dimensionality reduction techniques like PCA may not be sufficiently flexible to enable low-dimensional representation. Non-linear dimension reduction techniques, like kernel PCA and autoencoders, suffer from loss of interpretability since each latent variable is dependent of all input dimensions. To address this limitation, we here present path lasso penalized autoencoders. This structured regularization enhances interpretability by penalizing each path through the encoder from an input to a latent variable, thus restricting how many input variables are represented in each latent dimension. Our algorithm uses a group lasso penalty and non-negative matrix factorization to construct a sparse, non-linear latent representation. We compare the path lasso regularized autoencoder to PCA, sparse PCA, autoencoders and sparse autoencoders on real and simulated data sets. We show that the algorithm exhibits much lower reconstruction errors than sparse PCA and parameter-wise lasso regularized autoencoders for low-dimensional representations. Moreover, path lasso representations provide a more accurate reconstruction match, i.e. preserved relative distance between objects in the original and reconstructed spaces.

  • 38.
    Almér, Stefan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Control and Analysis of Pulse-Modulated Systems2008Doctoral thesis, comprehensive summary (Other scientific)
    Abstract [en]

    The thesis consists of an introduction and four appended papers. In the introduction we give an overview of pulse-modulated systems and provide a few examples of such systems. Furthermore, we introduce the so-called dynamic phasor model which is used as a basis for analysis in two of the appended papers. We also introduce the harmonic transfer function and finally we provide a summary of the appended papers.

    The first paper considers stability analysis of a class of pulse-width modulated systems based on a discrete time model. The systems considered typically have periodic solutions. Stability of a periodic solution is equivalent to stability of a fixed point of a discrete time model of the system dynamics.

    Conditions for global and local exponential stability of the discrete time model are derived using quadratic and piecewise quadratic Lyapunov functions. A griding procedure is used to develop a systematic method to search for the Lyapunov functions.

    The second paper considers the dynamic phasor model as a tool for stability analysis of a general class of pulse-modulated systems. The analysis covers both linear time periodic systems and systems where the pulse modulation is controlled by feedback. The dynamic phasor model provides an $\textbf{L}_2$-equivalent description of the system dynamics in terms of an infinite dimensional dynamic system. The infinite dimensional phasor system is approximated via a skew truncation. The truncated system is used to derive a systematic method to compute time periodic quadratic Lyapunov functions.

    The third paper considers the dynamic phasor model as a tool for harmonic analysis of a class of pulse-width modulated systems. The analysis covers both linear time periodic systems and non-periodic systems where the switching is controlled by feedback. As in the second paper of the thesis, we represent the switching system using the L_2-equivalent infinite dimensional system provided by the phasor model. It is shown that there is a connection between the dynamic phasor model and the harmonic transfer function of a linear time periodic system and this connection is used to extend the notion of harmonic transfer function to describe periodic solutions of non-periodic systems. The infinite dimensional phasor system is approximated via a square truncation. We assume that the response of the truncated system to a periodic disturbance is also periodic and we consider the corresponding harmonic balance equations. An approximate solution of these equations is stated in terms of a harmonic transfer function which is analogous to the harmonic transfer function of a linear time periodic system. The aforementioned assumption is proved to hold for small disturbances by proving the existence of a solution to a fixed point equation. The proof implies that for small disturbances, the approximation is good.

    Finally, the fourth paper considers control synthesis for switched mode DC-DC converters. The synthesis is based on a sampled data model of the system dynamics. The sampled data model gives an exact description of the converter state at the switching instances, but also includes a lifted signal which represents the inter-sampling behavior. Within the sampled data framework we consider H-infinity control design to achieve robustness to disturbances and load variations. The suggested controller is applied to two benchmark examples; a step-down and a step-up converter. Performance is verified in both simulations and in experiments.

    Download full text (pdf)
    FULLTEXT01
  • 39.
    Almér, Stefan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Sampled data control of DC-DC convertersArticle in journal (Other academic)
  • 40. Almér, Stefan
    et al.
    Jönsson, Ulf
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Dynamic Phasor Analysis Of Pulse-Modulated Systems2012In: SIAM Journal of Control and Optimization, ISSN 0363-0129, E-ISSN 1095-7138, Vol. 50, no 3, p. 1110-1138Article in journal (Refereed)
    Abstract [en]

    This paper considers stability and harmonic analysis of a general class of pulse-modulated systems. The systems are modeled using the dynamic phasor model, which explores the cyclic nature of the modulation functions by representing the system state as a Fourier series expansion defined over a moving time window. The contribution of the paper is to show that a special type of periodic Lyapunov function can be used to analyze the system and that the analysis conditions become tractable for computation after truncation. The approach provides a trade-off between complexity and accuracy that includes standard state space averaged models as a special case. The paper also shows how the dynamic phasor model can be used to derive a frequency domain input-to-state map which is analogous to the harmonic transfer function.

  • 41.
    Almér, Stefan
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Jönsson, Ulf
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Harmonic analysis of pulse-width modulated systems2009In: Automatica, ISSN 0005-1098, E-ISSN 1873-2836, Vol. 45, no 4, p. 851-862Article in journal (Refereed)
    Abstract [en]

    The paper considers the so-called dynamic phasor model as a basis for harmonic analysis of a class switching systems. The analysis covers both periodically switched systems and non-periodic systems where the switching is controlled by feedback. The dynamic phasor model is a powerful tool for exploring cyclic properties of dynamic systems. It is shown that there is a connection between the dynamic phasor model and the harmonic transfer function of a linear time periodic system and this connection is used to extend the notion of harmonic transfer function to describe periodic solutions of non-periodic systems.

  • 42.
    Almér, Stefan
    et al.
    KTH, Superseded Departments (pre-2005), Mathematics.
    Jönsson, Ulf
    KTH, Superseded Departments (pre-2005), Mathematics.
    Kao, Chung-Yao
    KTH, Superseded Departments (pre-2005), Mathematics.
    Mari, Jorge
    Global stability analysis of DC-DC converters using sampled-data modeling2004In: PROCEEDINGS OF THE 2004 AMERICAN CONTROL CONFERENCE, VOLS 1-6, 2004, p. 4549-4554Conference paper (Refereed)
    Abstract [en]

    The paper presents stability analysis of a class of pulse-width modulated (PWM) systems which incorporates many different DC-DC converters. Two types of pulse-width modulation (digital and analog control) are considered. A procedure is developed for systematic search for Lyapunov functions. The state space is partitioned in such a way that stability is verified if a set of coupled Linear Matrix Inequalities (LMIs) is feasible. Global stability is considered as well as the computation of local regions of attraction.

  • 43.
    Almér, Stefan
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Jönsson, Ulf
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Kao, Chung-Yao
    Univ Melbourne, Dept Elect & Elect Engn.
    Mari, Jorge
    GE Global Res, Elect Energy Syst.
    Stability analysis of a class of PWM systems2007In: IEEE Transactions on Automatic Control, ISSN 0018-9286, E-ISSN 1558-2523, Vol. 52, no 6, p. 1072-1078Article in journal (Refereed)
    Abstract [en]

    This note considers stability analysis of a class of pulsewidth modulated (PWM) systems that incorporates several different switched mode dc-de- converters. The systems of the class typically have periodic solutions. A sampled data model is developed and used to prove stability of these solutions. Conditions for global and local exponential stability are derived using quadratic and piecewise quadratic Lyapunov functions. The state space is partitioned and the stability conditions are verified by checking a set of coupled linear matrix inequalities (LMIs).

  • 44.
    Alpsten, Erik
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Modeling News Data Flows using Multivariate Hawkes Processes2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    This thesis presents a multivariate Hawkes process approach to model flows of news data. The data is divided into classes based on the news' content and sentiment levels, such that each class contains a homogeneous type of observations. The arrival times of news in each class are related to a unique element in the multivariate Hawkes process. Given this framework, the massive and complex flow of information is given a more compact representation that describes the excitation connections between news classes, which in turn can be used to better predict the future flow of news data. Such a model has potential applications in areas such as finance and security. This thesis focuses especially on the different bucket sizes used in the discretization of the time scale as well as the differences in results that these imply. The study uses aggregated news data provided by RavenPack and software implementations are written in Python using the TensorFlow package.

    For the cases with larger bucket sizes and datasets containing a larger number of observations, the results suggest that the Hawkes models give a better fit to training data than the Poisson model alternatives. The Poisson models tend to give better performance when models trained on historic data are tested on subsequent data flows. Moreover, the connections between news classes are given to vary significantly depending on the underlying datasets. The results indicate that lack of observations in certain news classes lead to over-fitting in the training of the Hawkes models and that the model ought to be extended to take into account the deterministic and periodic behaviors of the news data flows.

    Download full text (pdf)
    fulltext
  • 45.
    Alpsten, Gustav
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Samanci, Sercan
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Portfolio Protection Strategies: A study on the protective put and its extensions2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The need among investors to manage volatility has made itself painfully clear over the past century, particularly during sudden crashes and prolonged drawdowns in the global equity markets. This has given rise to a liquid portfolio insurance market in the form of options, as well as attracted the attention of many researchers. Previous literature has, in particular, studied the effectiveness of the widely known protective put strategy, which serially buys a put option to protect a long position in the underlying asset. The results are often uninspiring, pointing towards few, if any, protective benefits with high option premiums as a main concern. This raises the question if there are ways to improve the protective put strategy or if there are any cost-efficient alternatives that provide a relatively be.er protection. This study extends the previous literature by investigating potential improvements and alternatives to the protective put strategy. In particular, three alternative put spread strategies and one collar strategy are constructed. In addition, a modified protective put this introduced to mitigate the path dependency in a rolling protection strategy.

     

    The results show that no option-based protection strategy can dominate the other in all market situations. Although reducing the equity position is generally more effective than buying options, we report that a collar strategy that buys 5% OTM put options and sells 5% OTM call options has an attractive risk-reward profile and protection against drawdowns. We also show that the protective put becomes more effective, both in terms of risk-adjusted return and tail protection, for longer maturities.

    Download full text (pdf)
    fulltext
  • 46.
    Altmann, Robert
    et al.
    Univ Augsburg, Dept Math, Univ Str 14, D-86159 Augsburg, Germany..
    Henning, Patrick
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA. Ruhr Univ Bochum, Fac Math, D-44801 Bochum, Germany..
    Peterseim, Daniel
    Univ Augsburg, Dept Math, Univ Str 14, D-86159 Augsburg, Germany..
    Localization And Delocalization Of Ground States Of Bose-Einstein Condensates Under Disorder2022In: SIAM Journal on Applied Mathematics, ISSN 0036-1399, E-ISSN 1095-712X, Vol. 82, no 1, p. 330-358Article in journal (Refereed)
    Abstract [en]

    This paper studies the localization behavior of Bose-Einstein condensates in disorder potentials, modeled by a Gross-Pitaevskii eigenvalue problem on a bounded interval. In the regime of weak particle interaction, we are able to quantify exponential localization of the ground state, depending on statistical parameters and the strength of the potential. Numerical studies further show delocalization if we leave the identified parameter range, which is in agreement with experimental data. These mathematical and numerical findings allow the prediction of physically relevant regimes where localization of ground states may be observed experimentally.

  • 47.
    Altmann, Robert
    et al.
    Univ Augsburg, Inst Math, D-86159 Augsburg, Germany..
    Henning, Patrick
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA. Ruhr Univ Bochum, Fak Math, D-44801 Bochum, Germany..
    Peterseim, Daniel
    Univ Augsburg, Inst Math, D-86159 Augsburg, Germany..
    Numerical homogenization beyond scale separation2021In: Acta Numerica, ISSN 0962-4929, E-ISSN 1474-0508, Vol. 30, p. 1-86, article id PII S0962492921000015Article in journal (Refereed)
    Abstract [en]

    Numerical homogenization is a methodology for the computational solution of multiscale partial differential equations. It aims at reducing complex large-scale problems to simplified numerical models valid on some target scale of interest, thereby accounting for the impact of features on smaller scales that are otherwise not resolved. While constructive approaches in the mathematical theory of homogenization are restricted to problems with a clear scale separation, modern numerical homogenization methods can accurately handle problems with a continuum of scales. This paper reviews such approaches embedded in a historical context and provides a unified variational framework for their design and numerical analysis. Apart from prototypical elliptic model problems, the class of partial differential equations covered here includes wave scattering in heterogeneous media and serves as a template for more general multi-physics problems.

  • 48.
    Altmann, Robert
    et al.
    Univ Augsburg, Dept Math, Univ Str 14, D-86159 Augsburg, Germany..
    Henning, Patrick
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Numerical Analysis, NA. Ruhr Univ Bochum, Dept Math, D-44801 Bochum, Germany..
    Peterseim, Daniel
    Univ Augsburg, Dept Math, Univ Str 14, D-86159 Augsburg, Germany..
    The J-method for the Gross-Pitaevskii eigenvalue problem2021In: Numerische Mathematik, ISSN 0029-599X, E-ISSN 0945-3245, Vol. 148, no 3, p. 575-610Article in journal (Refereed)
    Abstract [en]

    This paper studies the J-method of [E. Jarlebring, S. Kvaal, W. Michiels. SIAM J. Sci. Comput. 36-4:A1978-A2001, 2014] for nonlinear eigenvector problems in a general Hilbert space framework. This is the basis for variational discretization techniques and a mesh-independent numerical analysis. A simple modification of the method mimics an energy-decreasing discrete gradient flow. In the case of the Gross-Pitaevskii eigenvalue problem, we prove global convergence towards an eigenfunction for a damped version of the J-method. More importantly, when the iterations are sufficiently close to an eigenfunction, the damping can be switched off and we recover a local linear convergence rate previously known from the discrete setting. This quantitative convergence analysis is closely connected to the J-method's unique feature of sensitivity with respect to spectral shifts. Contrary to classical gradient flows, this allows both the selective approximation of excited states as well as the amplification of convergence beyond linear rates in the spirit of the Rayleigh quotient iteration for linear eigenvalue problems. These advantageous convergence properties are demonstrated in a series of numerical experiments involving exponentially localized states under disorder potentials and vortex lattices in rotating traps.

  • 49.
    Alvianto Priyanto, Criss
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    Shift Design and Driver Scheduling Problem2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Scheduling problem and shift design problems are well known NP-hard problems within the optimization area. Often time, the two problems are studied individually. In this thesis however, we are looking at the combination of both problems. More specifically, the aim of this thesis is to suggest an optimal scheduling policy given that there are no predefined shifts to begin with. The duration of a shift, along with the start and end time may vary. Thus we have proposed to split the problem into two sub-problems: weekly scheduling problem and daily scheduling problem. As there are no exact solution methods that are feasible, two meta-heuristics method has been employed to solve the sub-problems: Simulated Annealing (SA) and Genetic Algorithm (GA). We have provided proofs of concepts for both methods as well as explored the scalability. This is especially important as the number of employee is expected to grow significantly throughout the year. The results obtained has shown to be promising and can be built upon for further capabilities.

    Download full text (pdf)
    fulltext
  • 50. Améndola, Carlos
    et al.
    Kohn, Kathlén
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Lamboglia, Sara
    Maclagan, Diane
    Smith, Ben
    Sommars, Jeff
    Tripoli, Paolo
    Zajaczkowska, Magdalena
    Computing Tropical Varieties in Macaulay2: arxiv.org:1710.10651Manuscript (preprint) (Other academic)
1234567 1 - 50 of 1234
CiteExportLink to result list
Permanent link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf