Adaptive Markov chain Monte Carlo and Bayesian filtering for state space models

This thesis is concerned with the state and parameter estimation in state space models. The estimation of states and parameters is an important task when mathematical modeling is applied to many different application areas such as the global positioning systems, target tracking, navigation, brain imaging, spread of infectious diseases, biological processes, telecommunications, audio signal processing, stochastic optimal control, machine learning, and physical systems. In Bayesian settings, the estimation of states or parameters amounts to computation of the posterior probability density function. Except for a very restricted number of models, it is impossible to compute this density function in a closed form. Hence, we need approximation methods. A state estimation problem involves estimating the states (latent variables) that are not directly observed in the output of the system. In this thesis, we use the Kalman filter, extended Kalman filter, Gauss–Hermite filters, and particle filters to estimate the states based on available measurements. Among these filters, particle filters are numerical methods for approximating the filtering distributions of non-linear non-Gaussian state space models via Monte Carlo. The performance of a particle filter heavily depends on the chosen importance distribution. For instance, inappropriate choice of the importance distribution can lead to the failure of convergence of the particle filter algorithm. In this thesis, we analyze the theoretical Lᵖ particle filter convergence with general importance distributions, where p ≥2 is an integer. A parameter estimation problem is considered with inferring the model parameters from measurements. For high-dimensional complex models, estimation of parameters can be done by Markov chain Monte Carlo (MCMC) methods. In its operation, the MCMC method requires the unnormalized posterior distribution of the parameters and a proposal distribution. In this thesis, we show how the posterior density function of the parameters of a state space model can be computed by filtering based methods, where the states are integrated out. This type of computation is then applied to estimate parameters of stochastic differential equations. Furthermore, we compute the partial derivatives of the log-posterior density function and use the hybrid Monte Carlo and scaled conjugate gradient methods to infer the parameters of stochastic differential equations. The computational efficiency of MCMC methods is highly depend on the chosen proposal distribution. A commonly used proposal distribution is Gaussian. In this kind of proposal, the covariance matrix must be well tuned. To tune it, adaptive MCMC methods can be used. In this thesis, we propose a new way of updating the covariance matrix using the variational Bayesian adaptive Kalman filter algorithm.

Fluvio-morphological processes of meander bends - Combining conventional field measurements, close-range remote sensing and computational modelling

Meandering rivers have been perceived to evolve rather similarly around the world independently of the location or size of the river. Despite the many consistent processes and characteristics they have also been noted to show complex and unique sets of fluviomorphological processes in which local factors play important role. These complex interactions of flow and morphology affect notably the development of the river. Comprehensive and fundamental field, flume and theoretically based studies of fluviomorphological processes in meandering rivers have been carried out especially during the latter part of the 20th century. However, as these studies have been carried out with traditional field measurements techniques their spatial and temporal resolution is not competitive to the level achievable today. The hypothesis of this study is that, by exploiting e increased spatial and temporal resolution of the data, achieved by combining conventional field measurements with a range of modern technologies, will provide new insights to the spatial patterns of the flow-sediment interaction in meandering streams, which have perceived to show notable variation in space and time. This thesis shows how the modern technologies can be combined to derive very high spatial and temporal resolution data on fluvio-morphological processes over meander bends. The flow structure over the bends is recorded in situ using acoustic Doppler current profiler (ADCP) and the spatial and temporal resolution of the flow data is enhanced using 2D and 3D CFD over various meander bends. The CFD are also exploited to simulate sediment transport. Multi-temporal terrestrial laser scanning (TLS), mobile laser scanning (MLS) and echo sounding data are used to measure the flow-based changes and formations over meander bends and to build the computational models. The spatial patterns of erosion and deposition over meander bends are analysed relative to the measured and modelled flow field and sediment transport. The results are compared with the classic theories of the processes in meander bends. Mainly, the results of this study follow well the existing theories and results of previous studies. However, some new insights regarding to the spatial and temporal patterns of the flow-sediment interaction in a natural sand-bed meander bend are provided. The results of this study show the advantages of the rapid and detailed measurements techniques and the achieved spatial and temporal resolution provided by CFD, unachievable with field measurements. The thesis also discusses the limitations which remain in the measurement and modelling methods and in understanding of fluvial geomorphology of meander bends. Further, the hydro- and morphodynamic models’ sensitivity to user-defined parameters is tested, and the modelling results are assessed against detailed field measurement. The study is implemented in the meandering sub-Arctic Pulmanki River in Finland. The river is unregulated and sand-bed and major morphological changes occur annually on the meander point bars, which are inundated only during the snow-melt-induced spring floods. The outcome of this study applies to sandbed meandering rivers in regions where normally one significant flood event occurs annually, such as Arctic areas with snow-melt induced spring floods, and where the point bars of the meander bends are inundated only during the flood events.

Simulating the dynamics of silver market using computational market dynamics

Traditional econometric approaches in modeling the dynamics of equity and commodity markets, have, made great progress in the past decades. However, they assume rationality among the economic agents and and do not capture the dynamics that produce extreme events (black swans), due to deviation from the rationality assumption. The purpose of this study is to simulate the dynamics of silver markets by using the novel computational market dynamics approach. To this end, the daily data from the period of 1st March 2000 to 1st March 2013 of closing prices of spot silver prices has been simulated with the Jabłonska-Capasso-Morale(JCM) model. The Maximum Likelihood approach has been employed to calibrate the acquired data with JCM. Statistical analysis of the simulated series with respect to the actual one has been conducted to evaluate model performance. The model captures the animal spirits dynamics present in the data under evaluation well.

Korkeapainevesisumusuuttimen virtaustekninen laadunhallinta

Tämä työ vastaa tarpeeseen hallita korkeapainevesisumusuuttimen laatua virtausmekaniikan työkalujen avulla. Työssä tutkitaan suutinten testidatan lisäksi virtauksen käyttäytymistä suuttimen sisällä CFD-laskennan avulla. Virtausmallinnus tehdään Navier-Stokes –pohjaisella laskentamenetelmällä. Työn teoriaosassa käsitellään virtaustekniikkaa ja sen kehitystä yleisesti. Lisäksi esitetään suuttimen laskennassa käytettävää perusteoriaa sekä teknisiä ratkaisuja. Teoriaosassa käydään myös läpi laskennalliseen virtausmekaniikkaan (CFD-laskenta) liittyvää perusteoriaa. Tutkimusosiossa esitetään käsitellyt suutintestitulokset sekä mallinnetaan suutinvirtausta ajasta riippumattomaan virtauslaskentaan perustuvalla laskentamenetelmällä. Virtauslaskennassa käytetään OpenFOAM-laskentaohjelmiston SIMPLE-virtausratkaisijaa sekä k-omega SST –turbulenssimallia. Tehtiin virtausmallinnus kaikilla paineilla, joita suuttimen testauksessa myös todellisuudessa käytetään. Lisäksi selvitettiin mahdolliset kavitaatiokohdat suuttimessa ja suunniteltiin kavitaatiota ehkäisevä suutingeometria. Todettiin myös lämpötilan ja epäpuhtauksien vaikuttavan kavitaatioon sekä mallinnettiin lämpötilan vaikutusta. Luotiin malli, jolla suuttimen suunnitteluun liittyviin haasteisiin voidaan vastata numeerisella laskennalla.

Kalman filter for computational market dynamics

Kalman filter is a recursive mathematical power tool that plays an increasingly vital role in innumerable fields of study. The filter has been put to service in a multitude of studies involving both time series modelling and financial time series modelling. Modelling time series data in Computational Market Dynamics (CMD) can be accomplished using the Jablonska-Capasso-Morale (JCM) model. Maximum likelihood approach has always been utilised to estimate the parameters of the JCM model. The purpose of this study is to discover if the Kalman filter can be effectively utilized in CMD. Ensemble Kalman filter (EnKF), with 50 ensemble members, applied to US sugar prices spanning the period of January, 1960 to February, 2012 was employed for this work. The real data and Kalman filter trajectories showed no significant discrepancies, hence indicating satisfactory performance of the technique. Since only US sugar prices were utilized, it would be interesting to discover the nature of results if other data sets are employed.

Computational studies of deformation in stainless steel rings due to torch heating

A support ring of AISI 304L stainless steel that holds vertical, parallel wires arranged in a circle forming a cylinder is studied. The wires are attached to the ring with heat-induced shrinkage. When the ring is heated with a torch the heat affected zone tries to expand while the adjacent cool structure obstructs the expansion causing upsetting. During cooling, the ring shrinks smaller than its original size clamping the wires. The most important requirement for the ring is that it should be as round as possible and the deformations should occur as overall shrinkage in the ring diameter. A three-dimensional nonlinear transient sequential thermo-structural Abaqus model is used together with a Fortran code that enters the heat flux to each affected element. The local and overall deformations in one ring inflicted by the heating are studied with a small amount of inspection on residual stresses. A variety of different cases are chosen to be studied with the model constructed to provide directional knowledge; torch flux with the means of speed, location of the wires, heating location and structural factors. The decrease of heating speed increases heat flux that rises the temperature increasing shrinkage. In a single progressive heating uneven distribution of shrinkage appears to the start/end region that can be partially fixed with using speeded heating’s to strengthen the heating of that region. Location of the wires affect greatly to the caused shrinkage unlike heating location. The ring structure affects also greatly to the shrinkage; smaller diameter, bigger ring height, thinner thickness and greater number of wires increase shrinkage.

A novel polar-based human face recognition computational model

Motivated by a recently proposed biologically inspired face recognition approach, we investigated the relation between human behavior and a computational model based on Fourier-Bessel (FB) spatial patterns. We measured human recognition performance of FB filtered face images using an 8-alternative forced-choice method. Test stimuli were generated by converting the images from the spatial to the FB domain, filtering the resulting coefficients with a band-pass filter, and finally taking the inverse FB transformation of the filtered coefficients. The performance of the computational models was tested using a simulation of the psychophysical experiment. In the FB model, face images were first filtered by simulated V1- type neurons and later analyzed globally for their content of FB components. In general, there was a higher human contrast sensitivity to radially than to angularly filtered images, but both functions peaked at the 11.3-16 frequency interval. The FB-based model presented similar behavior with regard to peak position and relative sensitivity, but had a wider frequency band width and a narrower response range. The response pattern of two alternative models, based on local FB analysis and on raw luminance, strongly diverged from the human behavior patterns. These results suggest that human performance can be constrained by the type of information conveyed by polar patterns, and consequently that humans might use FB-like spatial patterns in face processing.

Structural and Computational Existence Results for Multidimensional Subshifts

Symbolic dynamics is a branch of mathematics that studies the structure of infinite sequences of symbols, or in the multidimensional case, infinite grids of symbols. Classes of such sequences and grids defined by collections of forbidden patterns are called subshifts, and subshifts of finite type are defined by finitely many forbidden patterns. The simplest examples of multidimensional subshifts are sets of Wang tilings, infinite arrangements of square tiles with colored edges, where adjacent edges must have the same color. Multidimensional symbolic dynamics has strong connections to computability theory, since most of the basic properties of subshifts cannot be recognized by computer programs, but are instead characterized by some higher-level notion of computability. This dissertation focuses on the structure of multidimensional subshifts, and the ways in which it relates to their computational properties. In the first part, we study the subpattern posets and Cantor-Bendixson ranks of countable subshifts of finite type, which can be seen as measures of their structural complexity. We show, by explicitly constructing subshifts with the desired properties, that both notions are essentially restricted only by computability conditions. In the second part of the dissertation, we study different methods of defining (classes of ) multidimensional subshifts, and how they relate to each other and existing methods. We present definitions that use monadic second-order logic, a more restricted kind of logical quantification called quantifier extension, and multi-headed finite state machines. Two of the definitions give rise to hierarchies of subshift classes, which are a priori infinite, but which we show to collapse into finitely many levels. The quantifier extension provides insight to the somewhat mysterious class of multidimensional sofic subshifts, since we prove a characterization for the class of subshifts that can extend a sofic subshift into a nonsofic one.

Scalable Feature Selection Applications for Genome-Wide Association Studies of Complex Diseases

Personalized medicine will revolutionize our capabilities to combat disease. Working toward this goal, a fundamental task is the deciphering of geneticvariants that are predictive of complex diseases. Modern studies, in the formof genome-wide association studies (GWAS) have afforded researchers with the opportunity to reveal new genotype-phenotype relationships through the extensive scanning of genetic variants. These studies typically contain over half a million genetic features for thousands of individuals. Examining this with methods other than univariate statistics is a challenging task requiring advanced algorithms that are scalable to the genome-wide level. In the future, next-generation sequencing studies (NGS) will contain an even larger number of common and rare variants. Machine learning-based feature selection algorithms have been shown to have the ability to effectively create predictive models for various genotype-phenotype relationships. This work explores the problem of selecting genetic variant subsets that are the most predictive of complex disease phenotypes through various feature selection methodologies, including filter, wrapper and embedded algorithms. The examined machine learning algorithms were demonstrated to not only be effective at predicting the disease phenotypes, but also doing so efficiently through the use of computational shortcuts. While much of the work was able to be run on high-end desktops, some work was further extended so that it could be implemented on parallel computers helping to assure that they will also scale to the NGS data sets. Further, these studies analyzed the relationships between various feature selection methods and demonstrated the need for careful testing when selecting an algorithm. It was shown that there is no universally optimal algorithm for variant selection in GWAS, but rather methodologies need to be selected based on the desired outcome, such as the number of features to be included in the prediction model. It was also demonstrated that without proper model validation, for example using nested cross-validation, the models can result in overly-optimistic prediction accuracies and decreased generalization ability. It is through the implementation and application of machine learning methods that one can extract predictive genotype–phenotype relationships and biological insights from genetic data sets.

Using Computational Fluid-Dynamics (CFD) for the evaluation of beer pasteurization: effect of orientation of cans

In-package pasteurization is the most used method for beer microbiological stabilization. The search for safer and better quality food has created a need to better understand the processes involved in producing it. However, little is known about the temperature and velocity profiles during the thermal processes of liquid foods in commercial packaging, which results in over-dimensioned processes to guarantee safety, decreasing the sensorial and nutritional characteristics of the product and increasing process costs. Simulations using Computational Fluid-Dynamics (CFD) have been used by various authors to evaluate those processes. The objective of the present paper was to evaluate the effect of packaging orientation in the pasteurization of beer in a commercial aluminum can using CFD. A heating process was simulated at 60 ºC up to 15 PUs (a conventional beer process, in which 1 Pasteurization Unit (PU) is equivalent to 1minute at 60 ºC). The temperature profile and convection current velocity along the process and the variation of the PUs were evaluated in relation to time considering the cans in the conventional, inverted, and horizontal positions. The temperature and velocity profiles were similar to those presented in the literature. The package position did not result in process improvement.

Computational fluid dynamics evaluation of liquid food thermal process in a brick shaped package

Food processes must ensure safety and high-quality products for a growing demand consumer creating the need for better knowledge of its unit operations. The Computational Fluid Dynamics (CFD) has been widely used for better understanding the food thermal processes, and it is one of the safest and most frequently used methods for food preservation. However, there is no single study in the literature describing thermal process of liquid foods in a brick shaped package. The present study evaluated such process and the influence of its orientation on the process lethality. It demonstrated the potential of using CFD to evaluate thermal processes of liquid foods and the importance of rheological characterization and convection in thermal processing of liquid foods. It also showed that packaging orientation does not result in different sterilization values during thermal process of the evaluated fluids in the brick shaped package.

Analysis of Phase Separator Design Criteria Using Computational Fluid Dynamics

Gravitational phase separation is a common unit operation found in most large-scale chemical processes. The need for phase separation can arise e.g. from product purification or protection of downstream equipment. In gravitational phase separation, the phases separate without the application of an external force. This is achieved in vessels where the flow velocity is lowered substantially compared to pipe flow. If the velocity is low enough, the denser phase settles towards the bottom of the vessel while the lighter phase rises. To find optimal configurations for gravitational phase separator vessels, several different geometrical and internal design features were evaluated based on simulations using OpenFOAM computational fluid dynamics (CFD) software. The studied features included inlet distributors, vessel dimensions, demister configurations and gas phase outlet configurations. Simulations were conducted as single phase steady state calculations. For comparison, additional simulations were performed as dynamic single and two-phase calculations. The steady state single phase calculations provided indications on preferred configurations for most above mentioned features. The results of the dynamic simulations supported the utilization of the computationally faster steady state model as a practical engineering tool. However, the two-phase model provides more truthful results especially with flows where a single phase does not determine the flow characteristics.

Computational Market Dynamics of Virtual Economies

The last two decades have provided a vast opportunity to live and explore the compulsive imaginary world or virtual world through massively multiplayer online role-playing games (MMORPGs). MMORPG gives a wide range of opportunities to its users to participate with multi-players on the same platform, to communicate and to do real time actions. There is a virtual economy in these games which is largely player-driven. In-game currency provides its users to build up their Avatars, to buy or sell the necessary goods to play, survive in the games and so on. As a part of virtual economies generated through EVE Online, this thesis mainly focuses on how the prices of the minerals in EVE Online behave by applying the Jabłonska- Capasso-Morale (JCM) mathematical simulation model. It is to verify up to what degree the model can reproduce the virtual economy behavior. The model is applied to buy and sell prices of two minerals namely, isogen and morphite. The simulation results demonstrate that JCM model ts reasonably well to the mineral prices, which lets us conclude that virtual economies behave similarly to the real ones.

Power and Performance Management Strategies on NoC-Based Multicore Platforms, Using DVFS, Adaptive Algorithm and Task Parallelism

This work presents synopsis of efficient strategies used in power managements for achieving the most economical power and energy consumption in multicore systems, FPGA and NoC Platforms. In this work, a practical approach was taken, in an effort to validate the significance of the proposed Adaptive Power Management Algorithm (APMA), proposed for system developed, for this thesis project. This system comprise arithmetic and logic unit, up and down counters, adder, state machine and multiplexer. The essence of carrying this project firstly, is to develop a system that will be used for this power management project. Secondly, to perform area and power synopsis of the system on these various scalable technology platforms, UMC 90nm nanotechnology 1.2v, UMC 90nm nanotechnology 1.32v and UMC 0.18 μmNanotechnology 1.80v, in order to examine the difference in area and power consumption of the system on the platforms. Thirdly, to explore various strategies that can be used to reducing system’s power consumption and to propose an adaptive power management algorithm that can be used to reduce the power consumption of the system. The strategies introduced in this work comprise Dynamic Voltage Frequency Scaling (DVFS) and task parallelism. After the system development, it was run on FPGA board, basically NoC Platforms and on these various technology platforms UMC 90nm nanotechnology1.2v, UMC 90nm nanotechnology 1.32v and UMC180 nm nanotechnology 1.80v, the system synthesis was successfully accomplished, the simulated result analysis shows that the system meets all functional requirements, the power consumption and the area utilization were recorded and analyzed in chapter 7 of this work. This work extensively reviewed various strategies for managing power consumption which were quantitative research works by many researchers and companies, it's a mixture of study analysis and experimented lab works, it condensed and presents the whole basic concepts of power management strategy from quality technical papers.

GIS-based approach for optimization of onshore wind park infrastructure alignment in Finland

Wind power is a rapidly developing, low-emission form of energy production. In Fin-land, the official objective is to increase wind power capacity from the current 1 005 MW up to 3 500–4 000 MW by 2025. By the end of April 2015, the total capacity of all wind power project being planned in Finland had surpassed 11 000 MW. As the amount of projects in Finland is record high, an increasing amount of infrastructure is also being planned and constructed. Traditionally, these planning operations are conducted using manual and labor-intensive work methods that are prone to subjectivity. This study introduces a GIS-based methodology for determining optimal paths to sup-port the planning of onshore wind park infrastructure alignment in Nordanå-Lövböle wind park located on the island of Kemiönsaari in Southwest Finland. The presented methodology utilizes a least-cost path (LCP) algorithm for searching of optimal paths within a high resolution real-world terrain dataset derived from airborne lidar scannings. In addition, planning data is used to provide a realistic planning framework for the anal-ysis. In order to produce realistic results, the physiographic and planning datasets are standardized and weighted according to qualitative suitability assessments by utilizing methods and practices offered by multi-criteria evaluation (MCE). The results are pre-sented as scenarios to correspond various different planning objectives. Finally, the methodology is documented by using tools of Business Process Management (BPM). The results show that the presented methodology can be effectively used to search and identify extensive, 20 to 35 kilometers long networks of paths that correspond to certain optimization objectives in the study area. The utilization of high-resolution terrain data produces a more objective and more detailed path alignment plan. This study demon-strates that the presented methodology can be practically applied to support a wind power infrastructure alignment planning process. The six-phase structure of the method-ology allows straightforward incorporation of different optimization objectives. The methodology responds well to combining quantitative and qualitative data. Additional-ly, the careful documentation presents an example of how the methodology can be eval-uated and developed as a business process. This thesis also shows that more emphasis on the research of algorithm-based, more objective methods for the planning of infrastruc-ture alignment is desirable, as technological development has only recently started to realize the potential of these computational methods.