Uncertainty analysis in the Model Web

This thesis provides a set of tools for managing uncertainty in Web-based models and workflows.To support the use of these tools, this thesis firstly provides a framework for exposing models through Web services. An introduction to uncertainty management, Web service interfaces,and workflow standards and technologies is given, with a particular focus on the geospatial domain.An existing specification for exposing geospatial models and processes, theWeb Processing Service (WPS), is critically reviewed. A processing service framework is presented as a solutionto usability issues with the WPS standard. The framework implements support for Simple ObjectAccess Protocol (SOAP), Web Service Description Language (WSDL) and JavaScript Object Notation (JSON), allowing models to be consumed by a variety of tools and software. Strategies for communicating with models from Web service interfaces are discussed, demonstrating the difficultly of exposing existing models on the Web. This thesis then reviews existing mechanisms for uncertainty management, with an emphasis on emulator methods for building efficient statistical surrogate models. A tool is developed to solve accessibility issues with such methods, by providing a Web-based user interface and backend to ease the process of building and integrating emulators. These tools, plus the processing service framework, are applied to a real case study as part of the UncertWeb project. The usability of the framework is proved with the implementation of aWeb-based workflow for predicting future crop yields in the UK, also demonstrating the abilities of the tools for emulator building and integration. Future directions for the development of the tools are discussed.

Living with uncertainty in the age of runtime models

Uncertainty can be defined as the difference between information that is represented in an executing system and the information that is both measurable and available about the system at a certain point in its life-time. A software system can be exposed to multiple sources of uncertainty produced by, for example, ambiguous requirements and unpredictable execution environments. A runtime model is a dynamic knowledge base that abstracts useful information about the system, its operational context and the extent to which the system meets its stakeholders' needs. A software system can successfully operate in multiple dynamic contexts by using runtime models that augment information available at design-time with information monitored at runtime. This chapter explores the role of runtime models as a means to cope with uncertainty. To this end, we introduce a well-suited terminology about models, runtime models and uncertainty and present a state-of-the-art summary on model-based techniques for addressing uncertainty both at development- and runtime. Using a case study about robot systems we discuss how current techniques and the MAPE-K loop can be used together to tackle uncertainty. Furthermore, we propose possible extensions of the MAPE-K loop architecture with runtime models to further handle uncertainty at runtime. The chapter concludes by identifying key challenges, and enabling technologies for using runtime models to address uncertainty, and also identifies closely related research communities that can foster ideas for resolving the challenges raised. © 2014 Springer International Publishing.

Risk and uncertainty in supply markets and sustainability-related innovation

tentially valuable innovations. In energy policy, much attention is given to analysing and incentivising customer demand, but new technologies also need new supply markets, to provide products and services to build, operate and maintain the innovative technology. This paper addresses the impact of supply constraints on the long-term viability of sustainability related innovations, using the case of energy from waste (EfW). Uncertainties in the pricing and availability of feedstock (i.e. waste) deter potential investors in EfW projects. We draw on prior supply management research to conceptualise the problem, and identify what steps might be taken to address it. Based on this analysis, we propose a research agenda aimed at purchasing and supply scholars and centred on the need to understand better how markets evolve and how stakeholders can (legitimately) influence the evolution of supply markets to support the adoption of sustainability related innovation. Within this broad case, specific themes are recommended for further investigation.

Paternal low protein diet affects adult offspring cardiovascular and metabolic function in mice

Although the association between maternal periconceptional diet and adult offspring health is well characterised, our understanding of the impact of paternal nutrition at the time of conception on offspring phenotype remains poorly defined. Therefore, we determined the effect of a paternal preconception low protein diet (LPD on adult offspring cardiovascular and metabolic health in mice. Male C57BL/6 mice were fed either normal protein diet (NPD; 18% casein or LPD (9% casein for 7 wk before mating. At birth, a reduced male-to-female ratio (P = 0.03 and increased male offspring weight (P = 0.009 were observed in litters from LPD compared with NPD stud males with no differences in mean litter size. LPD offspring were heavier than NPD offspring at 2 and 3 wk of age (P <0.02. However, no subsequent differences in body weight were observed. Adult male offspring derived from LPD studs developed relative hypotension (decreased by 9.2 mmHg and elevated heart rate (P <0.05, whereas both male and female offspring displayed vascular dysfunction and impaired glucose tolerance relative to NPD offspring. At cull (24 wk, LPD males had elevated adiposity (P = 0.04, reduced heart-to-body weight ratio (P = 0.04, and elevated circulating TNF-α levels (P = 0.015 compared with NPD males. Transcript expression in offspring heart and liver tissue was reduced for genes involved in calcium signaling (Adcy, Plcb, Prkcb and metabolism (Fto in LPD offspring (P <0.03. These novel data reveal the impact of suboptimal paternal nutrition on adult offspring cardiovascular and metabolic homeostasis, and provide some insight into the underlying regulatory mechanisms.

Fuzzy knowledge-based approach to treating uncertainty in inventory control

Inventory control in complex manufacturing environments encounters various sources of uncertainity and imprecision. This paper presents one fuzzy knowledge-based approach to solving the problem of order quantity determination, in the presence of uncertain demand, lead time and actual inventory level. Uncertain data are represented by fuzzy numbers, and vaguely defined relations between them are modeled by fuzzy if-then rules. The proposed representation and inference mechanism are verified using a large numbers of examples. The results of three representative cases are summarized. Finally a comparison between the developed fuzzy knowledge-based and traditional, probabilistic approaches is discussed.

Integrating OpenMI and Uncertweb:managing uncertainty in OpenMI models

OpenMI is a widely used standard allowing exchange of data between integrated models, which has mostly been applied to dynamic, deterministic models. Within the FP7 UncertWeb project we are developing mechanisms and tools to support the management of uncertainty in environmental models. In this paper we explore the integration of the UncertWeb framework with OpenMI, to assess the issues that arise when propagating uncertainty in OpenMI model compositions, and the degree of integration possible with UncertWeb tools. In particular we develop an uncertainty-enabled model for a simple Lotka-Volterra system with an interface conforming to the OpenMI standard, exploring uncertainty in the initial predator and prey levels, and the parameters of the model equations. We use the Elicitator tool developed within UncertWeb to identify the initial condition uncertainties, and show how these can be integrated, using UncertML, with simple Monte Carlo propagation mechanisms. The mediators we develop for OpenMI models are generic and produce standard Web services that expose the OpenMI models to a Web based framework. We discuss what further work is needed to allow a more complete system to be developed and show how this might be used practically.

Visualizing uncertainty in multi-spectral remotely sensed imagery

Error and uncertainty in remotely sensed data come from several sources, and can be increased or mitigated by the processing to which that data is subjected (e.g. resampling, atmospheric correction). Historically the effects of such uncertainty have only been considered overall and evaluated in a confusion matrix which becomes high-level meta-data, and so is commonly ignored. However, some of the sources of uncertainty can be explicity identified and modelled, and their effects (which often vary across space and time) visualized. Others can be considered overall, but their spatial effects can still be visualized. This process of visualization is of particular value for users who need to assess the importance of data uncertainty for their own practical applications. This paper describes a Java-based toolkit, which uses interactive and linked views to enable visualization of data uncertainty by a variety of means. This allows users to consider error and uncertainty as integral elements of image data, to be viewed and explored, rather than as labels or indices attached to the data. © 2002 Elsevier Science Ltd. All rights reserved.

Optimization of the Investment Portfolio in the Conditions of Uncertainty

Portfolio analysis exists, perhaps, as long, as people think about acceptance of rational decisions connected with use of the limited resources. However the occurrence moment of portfolio analysis can be dated precisely enough is having connected it with a publication of pioneer work of Harry Markovittz (Markovitz H. Portfolio Selection) in 1952. The model offered in this work, simple enough in essence, has allowed catching the basic features of the financial market, from the point of view of the investor, and has supplied the last with the tool for development of rational investment decisions. The central problem in Markovitz theory is the portfolio choice that is a set of operations. Thus in estimation, both separate operations and their portfolios two major factors are considered: profitableness and risk of operations and their portfolios. The risk thus receives a quantitative estimation. The account of mutual correlation dependences between profitablenesses of operations appears the essential moment in the theory. This account allows making effective diversification of portfolio, leading to essential decrease in risk of a portfolio in comparison with risk of the operations included in it. At last, the quantitative characteristic of the basic investment characteristics allows defining and solving a problem of a choice of an optimum portfolio in the form of a problem of quadratic optimization.

Uncertainty and Fuzzy Sets: Classifying the Situation

The so called “Plural Uncertainty Model” is considered, in which statistical, maxmin, interval and Fuzzy model of uncertainty are embedded. For the last case external and internal contradictions of the theory are investigated and the modified definition of the Fuzzy Sets is proposed to overcome the troubles of the classical variant of Fuzzy Subsets by L. Zadeh. The general variants of logit- and probit- regression are the model of the modified Fuzzy Sets. It is possible to say about observations within the modification of the theory. The conception of the “situation” is proposed within modified Fuzzy Theory and the classifying problem is considered. The algorithm of the classification for the situation is proposed being the analogue of the statistical MLM(maximum likelihood method). The example related possible observing the distribution from the collection of distribution is considered.

Towards the Problems of an Evaluation of Data Uncertainty in Decision Support Systems

The question of forming aim-oriented description of an object domain of decision support process is outlined. Two main problems of an estimation and evaluation of data and knowledge uncertainty in decision support systems – straight and reverse, are formulated. Three conditions being the formalized criteria of aimoriented constructing of input, internal and output spaces of some decision support system are proposed. Definitions of appeared and hidden data uncertainties on some measuring scale are given.

A world full of surprises:Bayesian theory of surprise to quantify degrees of uncertainty

In the specific area of software engineering (SE) for self-adaptive systems (SASs) there is a growing research awareness about the synergy between SE and artificial intelligence (AI). However, just few significant results have been published so far. In this paper, we propose a novel and formal Bayesian definition of surprise as the basis for quantitative analysis to measure degrees of uncertainty and deviations of self-adaptive systems from normal behavior. A surprise measures how observed data affects the models or assumptions of the world during runtime. The key idea is that a "surprising" event can be defined as one that causes a large divergence between the belief distributions prior to and posterior to the event occurring. In such a case the system may decide either to adapt accordingly or to flag that an abnormal situation is happening. In this paper, we discuss possible applications of Bayesian theory of surprise for the case of self-adaptive systems using Bayesian dynamic decision networks. Copyright © 2014 ACM.

Supply market uncertainty:exploring consequences and responses within sustainability transitions

Often it is commercial, not technological, factors which hinder the adoption of potentially valuable innovations. In energy policy, much attention is given to analysing and incentivising consumer demand for renewable energy, but new technologies may also need new supply markets, to provide products and services to build, operate and maintain the innovative technology. This paper addresses the impact of supply constraints on the long-term viability of sustainability related innovations, using the case of bioenergy from organic waste. Uncertainties in the pricing and availability of feedstock (i.e. waste) may generate market deadlock and deter potential investors. We draw on prior research to conceptualise the problem, and identify what steps might be taken to address it. We propose a research agenda aimed at purchasing and supply scholars and centred on the need to understand better the interplay between market evolution and supply uncertainty and 'market shaping' - how stakeholders can legitimately influence supply market evolution - to support the adoption of sustainability related innovation.

A mixture density network approach to modelling and exploiting uncertainty in nonlinear control problems

In nonlinear and stochastic control problems, learning an efficient feed-forward controller is not amenable to conventional neurocontrol methods. For these approaches, estimating and then incorporating uncertainty in the controller and feed-forward models can produce more robust control results. Here, we introduce 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. A nonlinear multi-variable system with different delays between the input-output pairs is used to demonstrate the successful application of the developed control algorithm. The proposed method is suitable for redundant control systems and allows us to model strongly non-Gaussian distributions of control signal as well as processes with hysteresis. © 2004 Elsevier Ltd. All rights reserved.