IT Technical reports
http://www.it.uu.se/research/publications/reports
Technical reports from the Department of Information Technology, Uppsala University, Sweden
20230223T15:09:27Z
Department of Information Technology, Uppsala University, Sweden
Copyright © 2005 Department of Information Technology, Uppsala University, Sweden
daily
12
19990101T00:00+01:00

Technical report 2023001: Preconditioning of Discrete State and ControlConstrained Optimal Control ConvectionDiffusion Problems
http://www.it.uu.se/research/publications/reports/2023001
20230201
Ivo Dravins and Maya Neytcheva
<b>Abstract:</b> We consider the iterative solution of algebraic systems, arising in optimal control problems, constrained by a partial differential equation, with additional box constraints on the state and the control variables, and sparsity imposed on the control. A nonsymmetric twobytwo block preconditioner is analysed and tested for a wide range of problem, regularization and discretization parameters. The constraint equation characterizes convectiondiffusion processes.

Technical report 2022009: Analyzing the Parameter Bias when an ARMAX Model is Fitted to NoiseCorrupted Data
http://www.it.uu.se/research/publications/reports/2022009
20221001
Torsten Söderström and Umberto Soverini
<b>Abstract:</b> When an ARMAX model is fitted to noisecorrupted data using the prediction error method, biased estimates are obtained. The bias is examined, with emphasis on the situation when the system is almost nonidentifiable. In contrast to the case of using an output error model, no general results on the size of the bias seem to apply.

Technical report 2022008: Analyzing the Parameter Bias when an Instrumental Variable Method is Used with NoiseCorrupted Data
http://www.it.uu.se/research/publications/reports/2022008
20221001
Torsten Söderström and Umberto Soverini
<b>Abstract:</b> When an output error model is fitted to data with noisecorrupted inputs using a prediction error method, a bias occurs. It was previously shown that the bias is of order O(1/delta) for a small polezero separation delta. These notes examine the same problem when an instrumental variable model is fitted. A similar result is shown to hold for the instrumental variable case.

Technical report 2022007: Faster Functional Warming with Cache Merging
http://www.it.uu.se/research/publications/reports/2022007
20220801
Gustaf Borgström, Christian Rohner, and David BlackSchaffer
<b>Abstract:</b> SMARTSlike sampled hardware simulation techniques achieve good accuracy by simulating many small portions of an application in detail. However, while this reduces the detailed simulation time, it results in extensive cache warming times, as each of the many simulation points requires warming the whole memory hierarchy. Adaptive Cache Warming reduces this time by iteratively increasing warming until achieving sufficient accuracy. Unfortunately, each time the warming increases, the previous warming must be redone, nearly doubling the required warming. We address rewarming by developing a technique to merge the cache states from the previous and additional warming iterations. We address rewarming by developing a technique to merge the cache states from the previous and additional warming iterations. We demonstrate our merging approach on multilevel LRU cache hierarchy and evaluate and address the introduced errors. By removing warming redundancy, we expect an ideal 2x warming speedup when using our Cache Merging solution together with Adaptive Cache Warming. Experiments show that Cache Merging delivers an average speedup of 1.44x, 1.84x, and 1.87x for 128kB, 2MB, and 8MB L2 caches, respectively, with 95percentile absolute IPC errors of only 0.029, 0.015, and 0.006, respectively. These results demonstrate that Cache Merging yields significantly higher simulation speed with minimal losses.

Technical report 2022006: A Robust MultiGoal Exploration Aided Tracking Policy
http://www.it.uu.se/research/publications/reports/2022006
20220601
Ruoqi Zhang, Per Mattsson, and Torbjörn Wigren
<b>Abstract:</b> Setpoint control aims at finding a policy that can track a set point that varies over time. Such control objectives are central in industry, yet multigoal Reinforcement Learning methods are typically evaluated on other environments. The paper therefore proposes the use of a combination of feedback based amplitude aided exploration, simulated ensemble model training, together with policy optimization also over integrated errors, to arrive at a trained multigoal policy that can be directly deployed to realworld nonlinear setpoint control systems. The claim is supported by experiments with a realworld nonlinear cascaded tank process and a simulated strongly nonlinear pHcontrol system.

Technical report 2022005: Consistency Study of a Reconstructed Genotype Probability Distribution via Clustered Bootstrapping in NORB Pooling Blocks
http://www.it.uu.se/research/publications/reports/2022005
20220601
Camille Clouard and Carl Nettelblad
<b>Abstract:</b> For applications with biallelic genetic markers, group testing techniques, synonymous to pooling techniques, are usually applied for decreasing the cost of largescale testing as e.g. when detecting carriers of rare genetic variants. In some configurations, the results of the grouped tests cannot be decoded and the pooled items are missing. Inference of these missing items can be performed with specific statistical methods that are for example related to the ExpectationMaximization algorithm. Pooling has also been applied for determining the genotype of markers in large populations. The particularity of full genotype data for diploid organisms in the context of group testing are the ternary outcomes (two homozygous genotypes and one heterozygous), as well as the distribution of these three outcomes in a population, which is often ruled by the HardyWeinberg Equilibrium and depends on the allele frequency in such situation. When using a nonoverlapping repeated block pooling design, the missing items are only observed in particular arrangements. Overall, a data set of pooled genotypes can be described as an inference problem in Missing Not At Random data with nonmonotone missingness patterns. This study presents a preliminary investigation of the consistency of various iterative methods estimating the most likely genotype probabilities of the missing items in pooled data. We use the KullbackLeibler divergence and the L2 distance between the genotype distribution computed from our estimates and a simulated empirical distribution as a measure of the distributional consistency.

Technical report 2022004: StageParallel Preconditioners for Implicit RungeKutta Methods of Arbitrary High Order. Linear problems
http://www.it.uu.se/research/publications/reports/2022004
20220401
Owe Axelsson, Ivo Dravins, and Maya Neytcheva
<b>Abstract:</b> Fully implicit RungeKutta methods offer the possibility to use high order accurate time discretization to match space discretization accuracy, an issue of significant importance for many large scale problems of current interest, where we may have fine space resolution with many millions of spatial degrees of freedom and long time intervals. In this work we consider strongly Astable implicit RungeKutta methods of arbitrary order of accuracy, based on Radau quadratures. For the arising large algebraic systems we introduce an efficient preconditioner, that allows for fully stageparallel solution. We analyse the spectrum of the corresponding preconditioned system and illustrate the performance of the solution method with numerical experiments using MPI. In this work we consider only linear problems.

Technical report 2022003: Implicit Summation by Parts Operators for Finite Difference Approximations of First and Second Derivatives
http://www.it.uu.se/research/publications/reports/2022003
20220101
Ken Mattsson and Ylva Ljungberg Rydin
<b>Abstract:</b> Implicit finite difference approximations are derived for both the first and second derivates. The boundary closures are based on the bandednorm summationbyparts framework and the boundary conditions are imposed using a weak (penalty) enforcement. Up to 8th order global convergence is achieved. The finite difference approximations lead to implicit ODE systems. Spectral resolution characteristics are achieved by proper tuning of the internal difference stencils. The accuracy and stability properties are demonstrated for linear hyperbolic problems in 1D and the 2D compressible Euler equations.

Technical report 2022002: MATLAB Software for Nonlinear and Delayed Recursive Identification  Revision 2
http://www.it.uu.se/research/publications/reports/2022002
20220101
Torbjörn Wigren
<b>Abstract:</b> This report is the user's manual for a package of MATLAB scripts and functions, developed for recursive prediction error identification of nonlinear state space systems. The identified state space model incorporates delay, which allows a treatment of general nonlinear networked identification, as well as of general nonlinear systems with delay. The core of the package is an implementation of two output error identification algorithms. The algorithms are based on a continuous time, structured black box state space model of a nonlinear system. The present revision adds a new algorithm, where also the output is determined via a parameterized measurement equation in the states and inputs. The software can only be run offline, i.e. no true real time operation is possible. The algorithms are however implemented so that true online operation can be obtained by extraction of the main algorithmic loop. The user must then provide the real time environment. The software package contains scripts and functions that allow the user to either input live measurements or to generate test data by simulation. The scripts and functions for the setup and execution of the identification algorithms are somewhat more general than what is described in the references. The functionality for display of results include scripts for plotting of e.g. data, parameters, prediction errors, eigenvalues and the condition number of the Hessian. The estimated model obtained at the end of a run can be simulated and the model output plotted, alone or together with the data used for identification. Model validation is supported by two methods apart from the display functionality. First, a calculation of the RPEM loss function can be performed, using parameters obtained at the end of an identification run. Secondly, the accuracy as a function of the output signal amplitude can be assessed.

Technical report 2022001: Sjuksköterskors upplevelse av att jobba med ITsystem: sammanfattning
http://www.it.uu.se/research/publications/reports/2022001
20220101
Diane Golay and Åsa Cajander
<b>Abstract:</b> Denna rapport sammanfattar resultaten från en kvalitativ studie om sjuksköterskors upplevelse av ITsystem på jobbet. De känslor och uppfattningar som sjuksköterskor upplevde i samband med ITanvändning på jobbet presenteras och implikationerna för design och implementering av ITsystem och ITstödda processer i sjukhusmiljö diskuteras.

Technical report 2021008: MATLAB Software for Recursive Identification and Scaling Using a Structured Nonlinear Blackbox Model  Revision 7
http://www.it.uu.se/research/publications/reports/2021008
20211201
Torbjörn Wigren
<b>Abstract:</b> This reports is intended as a users manual for a package of MATLAB scripts and functions, developed for recursive prediction error identification of nonlinear state space systems and nonlinear static systems. The core of the package is the implementation of three output error identification and scaling algorithms. The first algorithm is based on a continuous time, structured black box state space model of a nonlinear system. An RPEM algorithm for recursive identification of nonlinear static systems, that reuses the parameterization of the nonlinear ODE model, is also included in the software package. The present revision adds a third algorithm, where also the output is determined via a parameterized measurement equation in the states and inputs. The software can only be run offline, i.e. no true real time operation is possible. The algorithm is however implemented so that true online operation can be obtained by extraction of the main algorithmic loop. The user must then provide the real time environment. The software package contains scripts and functions that allow the user to either input live measurements or to generate test data by simulation. The scripts and functions for the setup and execution of the identification algorithms are somewhat more general than what is described in the references. There is e.g. support for automatic reinitiation of the algorithms using the parameters obtained at the end of a previous identification run. This allows for multiple runs through a set of data, something that is useful for data sets that are too short to allow complete convergence. The reinitiation step also allows the user to modify the degrees of the polynomial model structure and to specify terms that are to be excluded from the model. This makes it possible to iteratively refine the estimated model using multiple runs. The functionality for display of results include scripts for plotting of data, parameters, prediction errors, eigenvalues and the condition number of the Hessian. The estimated model obtained at the end of a run can be simulated and the model output plotted, alone or together with the data used for identification. Model validation is supported by two methods apart from the display functionality. First, calculation of the RPEM loss function can be performed, using parameters obtained at the end of an identification run. Secondly, the accuracy as a function of the output signal amplitude can be assessed.

Technical report 2021007: MatrixLess Eigensolver for Large Structured Matrices
http://www.it.uu.se/research/publications/reports/2021007
20211101
Giovanni Barbarino, Melker Claesson, SvenErik Ekström, Carlo Garoni, David Meadon, and Hendrik Speleers
<b>Abstract:</b> Sequences of structured matrices of increasing size arise in many scientific applications and especially in the numerical discretization of linear differential problems. We assume as a working hypothesis that the eigenvalues of a matrix Xn belonging to a sequence of this kind are given by a regular expansion. Based on this working hypothesis, which is illustrated to be plausible through numerical experiments, we propose an eigensolver for the computation of the eigenvalues of Xn for large n and we provide a theoretical analysis of its convergence. The eigensolver is called matrixless because it does not operate on the matrix Xn but on a few similar matrices of smaller size combined with an interpolationextrapolation strategy. Its performance is benchmarked on several numerical examples, with a special focus on matrices arising from the discretization of differential problems.

Technical report 2021006: When are ErrorsinVariables Aspects Particularly Important to Consider in System Identification?
http://www.it.uu.se/research/publications/reports/2021006
20210901
Torsten Söderström and Umberto Soverini
<b>Abstract:</b> When recorded signals are corrupted by noise on both input and output sides, all standard identification methods give biased parameter estimates, due to the presence of input noise. This report discusses in what situations such a bias is large and, consequently, when the errorsinvariables identification methods are to be preferred.

Technical report 2021005: MatrixLess Eigensolver for Large Structured Matrices
http://www.it.uu.se/research/publications/reports/2021005
20210801
Giovanni Barbarino, Melker Claesson, SvenErik Ekström, Carlo Garoni, and David Meadon
<b>Abstract:</b> Sequences of structured matrices of increasing size arise in many scientific applications and especially in the numerical discretization of linear differential problems. We assume as a working hypothesis that the eigenvalues of a matrix Xn belonging to a sequence of this kind are given by a regular expansion. Based on the working hypothesis, which is proved to be plausible through numerical experiments, we propose an eigensolver for the computation of the eigenvalues of Xn for large n. The performance of the eigensolverwhich is called matrixless because it does not operate on the matrix Xnis illustrated on several numerical examples, with a special focus on matrices arising from the discretization of differential problems, and turns out to be quite satisfactory in all cases. In a sense, this is an a posteriori proof of the reasonableness of the working hypothesis as well as a testimony of the fact that the spectra of large structured matrices are much more "regular" than one might expect.