Deterministic Uncertainty Research Articles

МЕТОД АВТОМАТИЗОВАНОЇ КЛАСИФІКАЦІЇ НАДЗВИЧАЙНИХ СИТУАЦІЙ ІЗ СУДНОМ В АКВАТОРІЇ МОРСЬКОГО ПОРТУ

The aim of the study is to develop a method that allows to improve the accuracy of classification of emergencies with a ship in the seaport area. To achieve the research objective, an improved method for automated classification of ship emergencies in the seaport area under conditions of deterministic uncertainty is proposed. The method consists of methods for formalising and processing knowledge on determining the classes of ship emergencies. The method of knowledge formalisation is based on the formation of a set of factors influencing the classification of ship emergencies in the seaport water area based on fuzzy binary relations of non-strict preference. The method of knowledge processing is based on the formation of productive rules for classifying emergencies with a ship in the seaport area according to the predicted or current values of the desired factors in a fuzzy formulation. The determination of the membership functions of several fuzzy variables to linguistic variables is based on the processing of expert data represented by a matrix of binary relations of the values of the membership function of the elements of the domain of determining linguistic variables. A comparative assessment of the classes of recognisable emergencies with a vessel in the seaport water area involves solving a multi-criteria optimisation problem using the hierarchy analysis method. As a mathematical model for determining the classes of emergency situations with a vessel in the seaport waters, the article substantiates a logical-linguistic productive hierarchical model. The process of determining the classes of emergency is described with the help of an algebraic model, which is closest to the linguistic description. In case of inexpediency of synthesis of products, it is proposed to use the method of fuzzy identification to reduce the number of product rules. The proposed method allows to improve the accuracy of classification of emergencies with a ship in the seaport area under conditions of deterministic uncertainty. Keywords: classification, maritime port, hierarchy analysis, fuzzy identification, logicallinguistic production hierarchical model, vessel, class of emergency situations, fuzzy variable, linguistic variable

Impact and cost-effectiveness of strategies to prevent respiratory syncytial virus (RSV) disease in Vietnam: A modelling study

BackgroundNew prevention strategies for respiratory syncytial virus (RSV) are emerging, but it is unclear if they will be cost-effective in low- and middle-income countries. We evaluated the potential impact and cost-effectiveness of two strategies to prevent RSV disease in young children in Vietnam. MethodsWe used a static cohort model with a finely disaggregated age structure (weeks of age <5 years) to calculate the RSV disease burden in Vietnam, with and without a single dose of maternal vaccine (RSVpreF, Pfizer) or of monoclonal antibody (Nirsevimab, Sanofi, Astra Zeneca). Each strategy was compared to no pharmaceutical intervention, and to each other. We assumed both strategies would be administered year round over a ten-year period. The primary outcome measure was the cost per disability-adjusted life year (DALY) averted, from a societal perspective. We ran probabilistic and deterministic uncertainty analyses. ResultsWith central input assumptions for RSVpreF vaccine ($25/dose, 69 % efficacy, 6 months protection) and Nirsevimab ($25/dose, 77 % efficacy, 5 months protection), both options had similar cost-effectiveness ($3442 versus $3367 per DALY averted) when compared separately to no pharmaceutical intervention. RSVpreF vaccine had a lower net cost than Nirsevimab (net discounted cost of $213 m versus $264 m) but prevented fewer RSV deaths (24 % versus 31 %). Our results were very sensitive to assumptions about the dose price, efficacy, and duration of protection. At $5/dose and a willingness-to-pay threshold of 0.5 times the national GDP per capita, both prevention strategies have the potential to be cost-effective. ConclusionsRSVpreF vaccine and Nirsevimab may be cost-effective in Vietnam if appropriately priced.

CONTROL OF BUNDLE OF TRAJECTORIES OF A LINEAR DISCRETE SYSTEM WITH FINITE SETS OF INITIAL AND FINAL STATES

The article proposes an analysis of the problem of control of trajectory bundle of a linear discrete system. The problem of trajectory bundle control arises in applications characterized by the deterministic uncertainty of the initial conditions of the system, control of a group of objects of a similar nature, etc. To such problems belongs, in particular, the control problem for a beam of charged particles. In the statement of the problem studied in the paper, the control function is scalar, the set of initial states and the set of final states contain a finite number of elements. At the same time, the system is completely controlled. The problem is to find a control that transfers the system from a set of initial conditions to a set of final states. The problem is to find a control that transfers the system from a set of initial conditions to a set of final states. In the article, the problem of controlling a bundle of trajectories is reduced to the problem of terminal control of a linear discrete system of higher dimension. At the same time, the resulting system has a block structure. This approach is new. Applying the formula that establishes the dependence between the initial and final states of a discrete system, the problem of terminal control is reduced to the problem of finding a solution to a system of linear algebraic equations. Using the structure of the matrix of the system, as well as the form of the general solution of the system of linear algebraic equations, we obtain a general solution of the terminal control problem. Applying the properties of the pseudoinverse matrix, we prove the theorem on conditions of the solution existence and establish the function, which gives the general solution of the trajectory bundle control problem. This result has an algorithmic orientation.

Probabilistic analysis of wind turbine performance degradation due to blade erosion accounting for uncertainty of damage geometry

Geometry alterations of wind turbine blades due to erosion reduce the blade aerodynamic performance, yielding turbine power and energy losses. This study proposes a novel probabilistic analysis framework combining computational fluid dynamics, probabilistic and deterministic uncertainty propagation, and high-performance computing to assess this performance degradation accounting for the unavoidable uncertainty on field records of blade erosion. This uncertainty presently prevents using erosion records for improving wind turbine maintenance planning, increasing energy yield, and thus further reducing the wind energy cost. The technology is demonstrated by quantifying the statistical moments of power and energy yield losses of an eroded utility-scale turbine at a North Sea offshore site and a southern European onshore site. The expectations of the offshore and onshore annual energy production losses are found to be 2 and 3% of the corresponding nominal values, respectively, with corresponding standard deviations of 0.1 and 0.15%. In the realistic scenario of erosion varying with high radial frequency, these low standard deviations result from partial compensation of the impact of mild and severe damages. These low standard deviations indicate that present uncertainty levels of erosion geometry records can be handled with uncertainty analysis in predictive maintenance for further reducing wind energy costs. With the frequent assumption of small or no radial variation of erosion, the standard deviation of the loss is misleadingly higher. For the first time, the study reports on the significant impact of turbulence intensity of the installation site on the turbine loss variability with the site wind characteristics.

Depletion Perturbation Theory in decay heat calculation context

This article describes an implementation of Depletion Perturbation Theory (DPT) using deterministic neutronics code APOLLO3® alongside depletion code MENDEL. We use DPT to calculate the sensitivity coefficients of decay heat (DH) to various nuclear data such as reaction cross-sections, fissions yields, and reaction energy. DPT allows for the coupling of the Bateman, Boltzmann, and normalization equations, and has the advantage of needing only one calculation per cooling time to compute the sensitivity coefficient to all input parameters, as opposed to usual finite differences schemes which need as many calculation as there are input parameters. The calculation of sensitivity coefficients is done using both the DPT methodology and a finite differences scheme as a mean of numerical verification. Both methods showed good agreement, although we found that DPT should be used with care when studying short cooling times. Sensitivity coefficients were then used to perform deterministic uncertainty propagation, the resulting uncertainty calculation were used to compare the effect of various coupling configurations.

Adaptive Neural Task Space Control for Robot Manipulators With Unknown and Closed Control Architecture Under Random Vibrations

Robot manipulators are now used in various domains and environments, where they can be subjected to random vibrations. Random vibrations mainly affect the torque control signal, and a torque controller is therefore required to be designed for stabilization purposes. However, for security or intellectual property protection reasons, most commercialized robots are manufactured with unknown and inaccessible torque controller interface such that the user can only design a position/velocity controller. This paper proposes an adaptive task-space velocity controller free from the inner controller&#x2019;s structure and exhibiting stochastic and deterministic disturbances rejection to deal with these issues. To deal with the unknown inner controller, the paper exploits the fact that most torque controllers use a velocity feedback term, and it considers the other terms as an unknown functions vector. To cope with random disturbances, it is demonstrated that the random excitation matrix can be linearly parameterized, and therefore, a direct adaptive method is constructed. Using radial basis function neural network (RBF NN), an indirect adaptive method is developed to cope with deterministic uncertainties. Through Lyapunov theory, the paper proves that all the closed-loop signals are bounded in probability. The effectiveness of the proposed approach is further demonstrated through simulation comparisons.

Kolmogorovʼ entropy to investigate variation of forest soil properties in the Czech Republic

A measure of variation in values of physico-chemical properties of different soil groups is influenced by deterministic uncertainty. In this study, we assessed variation level of soil horizon properties in the most wide-spread forest soil groups on territory of the Czech Republic in the period 1953–2010. The assessment of variation was based on the selection of potentially correlating grain-size and chemical properties and a regression analysis (p < 0.05) between normal values (1981–2010) and Kolmogorov (K-) entropy of the forest soil quantities. Cambisols (73%), Stagnosols (10%) and Leptosols (4%) were the most frequently sampled soil groups in the database of 15,287 pits from the state pilot forest surveys during the normal period. The K-entropies of clay, Ntot, CaO and MgO indicated the mean contents in the soil groups more significantly than the entropies of loam, pH and Al2O3. The low variable clay content differentiated the series of Cambisols-Podzols (CM-PZ) from other soil groups. On the contrary, the higher stable clay content differentiated the Luvisols-Retisols-Stagnosols (LV-RT-ST) series. The low stable clay content and the higher Ntot and CaO contents were characteristic of Fluvisols and Gleysols. The CM-PZ series varied from LV-RT-ST by more variable contents of finer grain particles, Ntot and CaO. Relationships between soil properties means and K-entropies suggest different development series.

Interpretable, calibrated neural networks for analysis and understanding of inelastic neutron scattering data

Deep neural networks (NNs) provide flexible frameworks for learning data representations and functions relating data to other properties and are often claimed to achieve ‘super-human’ performance in inferring relationships between input data and desired property. In the context of inelastic neutron scattering experiments, however, as in many other scientific scenarios, a number of issues arise: (i) scarcity of labelled experimental data, (ii) lack of uncertainty quantification on results, and (iii) lack of interpretability of the deep NNs. In this work we examine approaches to all three issues. We use simulated data to train a deep NN to distinguish between two possible magnetic exchange models of a half-doped manganite. We apply the recently developed deterministic uncertainty quantification method to provide error estimates for the classification, demonstrating in the process how important realistic representations of instrument resolution in the training data are for reliable estimates on experimental data. Finally we use class activation maps to determine which regions of the spectra are most important for the final classification result reached by the network.

H∞ Constrained Pareto Suboptimal Strategy for Stochastic LPV Time-Delay Systems

Not only in control problems, but also in dynamic games, several sources of performance degradation, such as model variation, deterministic and stochastic uncertainties and state delays, need to be considered. In this paper, we present an [Formula: see text] constrained Pareto suboptimal strategy for stochastic linear parameter-varying (LPV) time-delay systems involving multiple decision makers. The goal of developing the [Formula: see text] constrained Pareto suboptimal strategy set is to construct a memoryless state feedback strategy set, so that the closed-loop stochastic LPV system is stochastically mean-square stable. In the paper, the existence condition of the extended bounded real lemma is first established via linear matrix inequalities (LMIs). Then, a quadratic cost bound for cost performance is derived. Based on these preliminary results, sufficient conditions for the existence of such a strategy set under the [Formula: see text] constraint are derived by using cross-coupled bilinear matrix inequalities (BMIs). To determine the strategy set, a viscosity iterative scheme based on the LMIs is established to avoid the processing of BMIs. Finally, two numerical examples are presented to demonstrate the reliability and usefulness of the proposed method.

Experiment-based affect heuristic using fuzzy rules and Taguchi statistical method for tuning complex systems

Affect heuristic is a subjective feeling that enables human's quick and efficient decision-making in unstructured and uncertain environments. It is an experience-based subconscious process in which information is framed for quick interpretation. Here, we propose a hybrid Experiment-based Decision-making architecture based on Affect Heuristics (EDAH) that uses possibility and probability theories in combination utilizing fuzzy logic's ability to address deterministic uncertainty and graduality of truth, along with Taguchi method's ability to exploit the statistical data as a robust and suboptimal experiment design methodology. Accordingly, the proposed paradigm aims to make the best decision from a set of feasible decisions using fewer experiments by inserting reliable subjective information by the expert in the random samples of experiments. The proposed method is mathematically defined and then applied to the problem of Targeted Drug Delivery (TDD) for cancer treatment in the two states of with and without noise. Furthermore, we study the effect of three membership functions, consisting of triangular, trapezoidal, and Gaussian forms on our decisions' selection for making an improved medical decision and avoiding an unprofitable one by recognizing existing opportunities. Results indicate considerable improvement of normoxic and hypoxic cell densities when compared with the competing approaches.

Identifiability and physical interpretability of hybrid, gray-box models - a case study

Model identifiability concerns the uniqueness of uncertain model parameters to be estimated from available process data and is often thought of as a prerequisite for the physical interpretability of a model. Nevertheless, model identifiability may be challenging to obtain in practice due to both stochastic and deterministic uncertainties, e.g. low data variability, noisy measurements, erroneous model structure, and stochasticity and locality of the optimization algorithm. For gray-box, hybrid models, model identifiability is rarely obtainable due to a high number of parameters. We illustrate through an industrial case study - modeling of a production choke valve in a petroleum well - that physical interpretability may be preserved even for non-identifiable models with adequate parameter regularization in the estimation problem. To this end, in a real industrial scenario, it may be beneficial for the model's predictive performance to develop hybrid over mechanistic models, as the model flexibility is higher. Modeling of six petroleum wells on the asset Edvard Grieg using historical production data show a 35\% reduction in the median prediction error across the wells comparing a hybrid to a mechanistic model. On the other hand, both the predictive performance and physical interpretability of the developed models are influenced by the available data. The findings encourage research into online learning and other hybrid model variants to improve the results.

Improving Uncertainty Estimation With Semi-Supervised Deep Learning for COVID-19 Detection Using Chest X-Ray Images.

In this work we implement a COVID-19 infection detection system based on chest X-ray images with uncertainty estimation. Uncertainty estimation is vital for safe usage of computer aided diagnosis tools in medical applications. Model estimations with high uncertainty should be carefully analyzed by a trained radiologist. We aim to improve uncertainty estimations using unlabelled data through the MixMatch semi-supervised framework. We test popular uncertainty estimation approaches, comprising Softmax scores, Monte-Carlo dropout and deterministic uncertainty quantification. To compare the reliability of the uncertainty estimates, we propose the usage of the Jensen-Shannon distance between the uncertainty distributions of correct and incorrect estimations. This metric is statistically relevant, unlike most previously used metrics, which often ignore the distribution of the uncertainty estimations. Our test results show a significant improvement in uncertainty estimates when using unlabelled data. The best results are obtained with the use of the Monte Carlo dropout method.

Representing Model Discrepancy in Bound-to-Bound Data Collaboration

We extend the existing methodology in bound-to-bound data collaboration (B2BDC), an optimization-based deterministic uncertainty quantification (UQ) framework, to explicitly take into account model discrepancy. The discrepancy is represented as a linear combination of finite basis functions, and the feasible set is constructed according to a collection of modified model-data constraints. Formulas for making predictions are also modified to include the model discrepancy function. Prior information about the model discrepancy can be added to the framework as additional constraints. Dataset consistency, a central feature of B2BDC, is generalized based on the extended framework.

Nuclear Data Uncertainty Propagation in Complex Fusion Geometries

The ASUSD program package was designed to automate and simplify the process of deterministic nuclear data sensitivity and uncertainty quantification. The program package couples Denovo, a discrete ordinate 3D transport solver, as part of ADVANTG and SUSD3D, a deterministic first order perturbation theory based Sensitivity/Uncertainty code, using several auxiliary programs used for input data preparation and post processing. Because of the automation employed in ASUSD, it is useful for Sensitivity/Uncertainty analysis of complex fusion geometries. In this paper, ASUSD was used to quantify uncertainties in the JET KN2 irradiation position. The results were compared to previously obtained probabilistic-based uncertainties determined using TALYS-based random nuclear data samples and MCNP in a Total Monte Carlo computation scheme. Results of the two approaches, deterministic and probabilistic, to nuclear data uncertainty propagation are compared and discussed. ASUSD was also used to perform preliminary Sensitivity/Uncertainty (S/U) analyses of three JET3-NEXP streaming benchmark experimental positions (A1, A4 and A7).

Economic evaluation of the introduction of rotavirus vaccine in Hong Kong

BackgroundRotavirus is a common cause of severe gastroenteritis in young children in Hong Kong (HK) with a high economic burden. This study aimed to evaluate the cost-effectiveness of introducing rotavirus vaccination into the HK Government’s Childhood Immunisation Programme (CIP) and to include the potential protective effect of the vaccine against seizures. MethodsA decision-support model was customised to estimate the potential impact, cost-effectiveness and benefit-risk of rotavirus vaccination in children below 5 years over the period 2020–2029 in HK. Two doses of Rotarix® and three doses of RotaTeq® were each compared to no vaccination. Rotavirus treatment costs were calculated from a governmental health sector perspective (i.e., costs of public sector treatment) and an overall health sector perspective (both governmental and patient, i.e., costs of public sector treatment, private sector treatment, transport and diapers). We ran probabilistic and deterministic uncertainty analyses. ResultsIntroduction of rotavirus vaccination in HK could prevent 49,000 (95% uncertainty interval: ~44,000–54,000) hospitalisations of rotavirus gastroenteritis and seizures and result in ~50 (95% uncertainty interval: ~25–85) intussusception hospitalisations, over the period 2020–2029 (a benefit-risk ratio of ~1000:1), compared to a scenario with no public or private sector vaccine use. The discounted vaccination cost would be US$51–57 million over the period 2020–2029 based on per-course prices of US$72 (Rotarix®) or US$78 (RotaTeq®), but this would be offset by discounted treatment cost savings of US$70 million (government) and US$127 million (governmental and patient health sector). There was a greater than 94% probability that the vaccine could be cost-saving irrespective of the vaccine product or perspective considered. All deterministic ‘what-if’ scenarios were cost-saving from an overall health sector perspective (governmental and patient). ConclusionsRotavirus vaccination is likely to be cost-saving and have a favourable benefit-risk profile in HK. Based on the assumptions made, our analysis supports its introduction into CIP.

Distributed Robust Optimization Algorithms Over Uncertain Network Graphs.

This article investigates the robustness issues of a set of distributed optimization algorithms, which aim to approach the optimal solution to a sum of local cost functions over an uncertain network. The uncertain communication network consists of transmission channels perturbed by additive deterministic uncertainties, which can describe quantization and transmission errors. A new robust initialization-free algorithm is proposed for the distributed optimization problem of multiple Euler-Lagrange systems, and the explicit relationship of the feedback gain of the algorithm, the communication topology, the properties of the cost function, and the radius of the channel uncertainties is established in order to reach the optimal solution. This result provides a sufficient condition for the selection of the feedback gain when the uncertainty size is less than the unity. As a special case, we discuss the impact of communication uncertainties on the distributed optimization algorithms for first-order integrator networks.

Analysis of Runoff-Sediment Cointegration and Uncertainty Relations at Different Temporal Scales in the Middle Reaches of the Yellow River, China

To understand the intricate runoff-sediment relationship in the middle Yellow River basin (MYRB), the Toudaoguai, Longmen, Tongguan and Huayuankou sites in the MYRB were selected to analyze the deterministic equilibrium and uncertainty relations of runoff-sediment based on 55-year hydrological data at multi-time scales. The Johansen test and wavelet neural network were used to verify the cointegration relationship among hydrological series. Runoff-sediment uncertain statistical relations and dynamics in the MYRB were also analyzed based on rating curves and hysteresis loops. The results showed that the logarithmic sequences of sediment load (SL), runoff and suspended sediment concentration (SSC) conformed to a linear cointegration relationship at the Toudaoguai station or in spring, winter or under small flow at other stations, but a nonlinear cointegration relationship was observed in other cases at other stations. Regarding runoff-sediment uncertain relationships, the rating curves, and hysteresis loops differed in stations (Toudaoguai and the other stations), as well as discharge (threshold: 1000 m3/s), season (ice-flood and rainy season) and saturation of flow at flood and monthly scales. At the annual scale, phased and unsynchronized characteristics of runoff and sediment load were evident with a decreasing trend. This study on the runoff-sediment relationship can rationally provide a theoretical basis for the management and development of the Yellow River and other similar rivers with sufficient sediment, especially for areas with serious soil erosion.

Cost-effectiveness of introducing a domestic pneumococcal conjugate vaccine (PCV7-TT) into the Cuban national immunization programme

ObjectivesTo evaluate the cost-effectiveness of introducing a domestic pneumococcal conjugate vaccine (PCV7-TT) into the Cuban National Immunization Program (NIP). MethodsWe compared PCV7-TT given at two, four and six months of age to a scenario without PCV7-TT, over a ten-year period (2020–2029). We calculated the cost (Cuban pesos – CUP) per Disability Adjusted Life Year (DALY) averted from a Government perspective. We compared results from a static cohort model and a parsimonious prediction model informed by the serotype distribution among pneumococcal carriers and cases. We ran probabilistic and deterministic uncertainty analyses. ResultsPCV7-TT could prevent 6897 (95% uncertainty interval, 4344–8750) hospitalizations and 189 (115–253) deaths in children <5 years of age, over the period 2020–2029. This could cost around 25 million (20–31) discounted CUP but would be offset by treatment cost savings of around 23 million (14–31). A parsimonious model predicted less favourable impact and cost-effectiveness but the cost per DALY averted was still less than 0.4 times the current GDP per capita. ConclusionsPCV7-TT is likely to be cost-effective in Cuba. The impact of the vaccine would need to be carefully monitored following its introduction into the NIP.

Economic evaluation of health education programs: The case of adolescent reproductive health in Vietnam

ABSTRACT Objectives: To examine the cost-effectiveness of three mutually exclusive reproductive health interventions (named level A, B and C) implemented in Chi Linh for adolescents. Methods: This research involved the standard techniques of cost-utility analysis. A Markov state-transition model was used to show how a hypothetical cohort of 100,000 adolescents moved between 11 different health states over time, beginning at the year the intervention commenced (2011) and continuing in 3-month increments. The uncertainty of model input parameters was examined in probabilistic sensitivity analyses. Results: The deterministic analysis and uncertainty analysis revealed that implementing intervention level B compared to level A was highly cost effective. The ICER of intervention level B in relation to level A was USD 915/QALY gained. Implementing intervention level C compared to level B was cost-effective. The ICER of intervention level C over level B was USD 1911/QALY gained. Conclusions: If the decision makers including the intervention team, Ministry of Health and Ministry of Education and Training, change their current practice of educating adolescents on reproductive health to the best option, costs would be incurred at the lowest amount and health benefits would be brought back to the highest level.

Distributed state estimation for uncertain linear systems: A regularized least-squares approach

This paper addresses the state estimation problem for a discrete-time uncertain system with a network of sensors, where the system is not necessarily observable by each sensor and deterministic uncertainties exist in the system matrices. A new robust estimator is designed for each sensor, using only its own and neighbor’s information, which is fully distributed. Moreover, a novel information fusion strategy is developed to guarantee the estimation performance, based on the collective observability of the sensor network, which greatly relaxes the technical assumption of the proposed estimator. Theoretically, it can be ensured that if the observed system is time-varying, the gains of the estimator will be bounded. Furthermore, if the system is time-invariant, these gains will be convergent. Subsequently, the estimation error covariance will be ultimately bounded if the observed system is quadratically bounded. In the end, the superiority of the proposed robust distributed state estimation algorithm is illustrated by several numerical simulation examples.