Estimating parameters in stochastic systems:a variational Bayesian approach

This work is concerned with approximate inference in dynamical systems, from a variational Bayesian perspective. When modelling real world dynamical systems, stochastic differential equations appear as a natural choice, mainly because of their ability to model the noise of the system by adding a variation of some stochastic process to the deterministic dynamics. Hence, inference in such processes has drawn much attention. Here a new extended framework is derived that is based on a local polynomial approximation of a recently proposed variational Bayesian algorithm. The paper begins by showing that the new extension of this variational algorithm can be used for state estimation (smoothing) and converges to the original algorithm. However, the main focus is on estimating the (hyper-) parameters of these systems (i.e. drift parameters and diffusion coefficients). The new approach is validated on a range of different systems which vary in dimensionality and non-linearity. These are the Ornstein–Uhlenbeck process, the exact likelihood of which can be computed analytically, the univariate and highly non-linear, stochastic double well and the multivariate chaotic stochastic Lorenz ’63 (3D model). As a special case the algorithm is also applied to the 40 dimensional stochastic Lorenz ’96 system. In our investigation we compare this new approach with a variety of other well known methods, such as the hybrid Monte Carlo, dual unscented Kalman filter, full weak-constraint 4D-Var algorithm and analyse empirically their asymptotic behaviour as a function of observation density or length of time window increases. In particular we show that we are able to estimate parameters in both the drift (deterministic) and the diffusion (stochastic) part of the model evolution equations using our new methods.

Managing uncertainty in complex stochastic models: design and emulation of a Rabies model

In this paper we present a novel method for emulating a stochastic, or random output, computer model and show its application to a complex rabies model. The method is evaluated both in terms of accuracy and computational efficiency on synthetic data and the rabies model. We address the issue of experimental design and provide empirical evidence on the effectiveness of utilizing replicate model evaluations compared to a space-filling design. We employ the Mahalanobis error measure to validate the heteroscedastic Gaussian process based emulator predictions for both the mean and (co)variance. The emulator allows efficient screening to identify important model inputs and better understanding of the complex behaviour of the rabies model.

Have the GIPSI settled down? Breaks and multivariate stochastic volatility models for, and not against, the European financial integration

We investigate the integration of the European peripheral financial markets with Germany, France, and the UK using a combination of tests for structural breaks and return correlations derived from several multivariate stochastic volatility models. Our findings suggest that financial integration intensified in anticipation of the Euro, further strengthened by the EMU inception, and amplified in response to the 2007/2008 financial crisis. Hence, no evidence is found of decoupling of the equity markets in more troubled European countries from the core. Interestingly, the UK, despite staying outside the EMU, is not worse integrated with the GIPSI than Germany or France. © 2013 Elsevier B.V.

Methods of energy recovery in the iron and steel industry: evalution of new possibilities and design studies for slab cooling

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY WITH PRIOR ARRANGEMENT

Modelling nonlinear stochastic dynamics in financial time series

For analysing financial time series two main opposing viewpoints exist, either capital markets are completely stochastic and therefore prices follow a random walk, or they are deterministic and consequently predictable. For each of these views a great variety of tools exist with which it can be tried to confirm the hypotheses. Unfortunately, these methods are not well suited for dealing with data characterised in part by both paradigms. This thesis investigates these two approaches in order to model the behaviour of financial time series. In the deterministic framework methods are used to characterise the dimensionality of embedded financial data. The stochastic approach includes here an estimation of the unconditioned and conditional return distributions using parametric, non- and semi-parametric density estimation techniques. Finally, it will be shown how elements from these two approaches could be combined to achieve a more realistic model for financial time series.

Exploiting uncertainty in nonlinear stochastic control problem

This work introduces a novel inversion-based neurocontroller for solving control problems involving uncertain nonlinear systems which could also compensate for multi-valued systems. The approach uses recent developments in neural networks, especially in the context of modelling statistical distributions, which are applied to forward and inverse plant models. Provided that certain conditions are met, an estimate of the intrinsic uncertainty for the outputs of neural networks can be obtained using the statistical properties of networks. More generally, multicomponent distributions can be modelled by the mixture density network. Based on importance sampling from these distributions a novel robust inverse control approach is obtained. This importance sampling provides a structured and principled approach to constrain the complexity of the search space for the ideal control law. The developed methodology circumvents the dynamic programming problem by using the predicted neural network uncertainty to localise the possible control solutions to consider. Convergence of the output error for the proposed control method is verified by using a Lyapunov function. Several simulation examples are provided to demonstrate the efficiency of the developed control method. The manner in which such a method is extended to nonlinear multi-variable systems with different delays between the input-output pairs is considered and demonstrated through simulation examples.

Solidification and cooling of grey cast iron in shell moulds with particular reference to the sub-surface structure

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

Properties of seawater bitterns with regard to liquid-desiccant cooling

This study concerns the properties of the concentrated bittern solutions occurring as by-product from solar salt works, in relation to their potential use as liquid desiccants in cooling systems. Solutions of compositions similar to those of bitterns have been made up in the laboratory, as have concentrated mixtures of MgCl2–MgSO4–H2O. Measurements of vapour pressure have been carried out using an isoteniscope and are reported together with measurements of density and viscosity. Several theoretical models representing these properties are reviewed and compared against the experimental results; the average agreement between theory and experiment is within 5% for vapour pressure and better for the preferred models of the other two properties. Based on these findings, an expression is provided for the equilibrium relative humidity of bitterns as a function of concentration relative to raw seawater. The vapour pressures of bittern solutions are found to be similar to those of solutions containing only magnesium chloride but having the same mass fraction of total salts. Therefore magnesium chloride solution is a reasonable model for bitterns for the purpose of developing the proposed cooling system.

A Bayesian perspective on stochastic neurocontrol

Control design for stochastic uncertain nonlinear systems is traditionally based on minimizing the expected value of a suitably chosen loss function. Moreover, most control methods usually assume the certainty equivalence principle to simplify the problem and make it computationally tractable. We offer an improved probabilistic framework which is not constrained by these previous assumptions, and provides a more natural framework for incorporating and dealing with uncertainty. The focus of this paper is on developing this framework to obtain an optimal control law strategy using a fully probabilistic approach for information extraction from process data, which does not require detailed knowledge of system dynamics. Moreover, the proposed control method framework allows handling the problem of input-dependent noise. A basic paradigm is proposed and the resulting algorithm is discussed. The proposed probabilistic control method is for the general nonlinear class of discrete-time systems. It is demonstrated theoretically on the affine class. A nonlinear simulation example is also provided to validate theoretical development.

Stochastic beamforming for cochlear implant coding

Cochlear implants are prosthetic devices used to provide hearing to people who would otherwise be profoundly deaf. The deliberate addition of noise to the electrode signals could increase the amount of information transmitted, but standard cochlear implants do not replicate the noise characteristic of normal hearing because if noise is added in an uncontrolled manner with a limited number of electrodes then it will almost certainly lead to worse performance. Only if partially independent stochastic activity can be achieved in each nerve fibre can mechanisms like suprathreshold stochastic resonance be effective. We are investigating the use of stochastic beamforming to achieve greater independence. The strategy involves presenting each electrode with a linear combination of independent Gaussian noise sources. Because the cochlea is filled with conductive salt solutions, the noise currents from the electrodes interact and the effective stimulus for each nerve fibre will therefore be a different weighted sum of the noise sources. To some extent therefore, the effective stimulus for a nerve fibre will be independent of the effective stimulus of neighbouring fibres. For a particular patient, the electrode position and the amount of current spread are fixed. The objective is therefore to find the linear combination of noise sources that leads to the greatest independence between nerve discharges. In this theoretical study we show that it is possible to get one independent point of excitation (one null) for each electrode and that stochastic beamforming can greatly decrease the correlation between the noise exciting different regions of the cochlea. © 2007 Copyright SPIE - The International Society for Optical Engineering.

Concentrated seawater brines for use in solar-powered desiccant cooling cycles

We have directly measured properties of concentrated seawater brines produced through solar evaporation in salt works. They are sufficiently hygroscopic for use in desiccant cooling cycles which can cool air to 8.0–10.9 °C below ambient. This compares to only 3.8–8.7 °C with simple evaporative cooling. Desiccant cooling can extend the growing seasons of greenhouse crops thus providing an adaptive measure against climate change.

Modelling and experimental verification of a solar-powered liquid desiccant cooling system for greenhouse food production in hot climates

Experiments and theoretical modelling have been carried out to predict the performance of a solar-powered liquid desiccant cooling system for greenhouses. We have tested two components of the system in the laboratory using MgCl2 desiccant: (i) a regenerator which was tested under a solar simulator and (ii) a desiccator which was installed in a test duct. Theoretical models have been developed for both regenerator and desiccator and gave good agreement with the experiments. The verified computer model is used to predict the performance of the whole system during the hot summer months in Mumbai, Chittagong, Muscat, Messina and Havana. Taking examples of temperate, sub-tropical, tropical and heat-tolerant tropical crops (lettuce, soya bean, tomato and cucumber respectively) we estimate the extensions in growing seasons enabled by the system. Compared to conventional evaporative cooling, the desiccant system lowers average daily maximum temperatures in the hot season by 5.5-7.5 °C, sufficient to maintain viable growing conditions for lettuce throughout the year. In the case of tomato, cucumber and soya bean the system enables optimal cultivation through most summer months. It is concluded that the concept is technically viable and deserves testing by means of a pilot installation at an appropriate location.