Volumen 27 (2017): Heft 2 (June 2017)

This paper proposes a novel computationally efficient stochastic spectral projection based approach to Bayesian inversion of a computer simulator with high dimensional parametric and model structure uncertainty. The proposed method is based on the decomposition of the solution into its mean and a random field using a generic Karhunen-Loève expansion. The random field is represented as a convolution of separable Hilbert spaces in stochastic and spatial dimensions that are spectrally represented using respective orthogonal bases. In particular, the present paper investigates generalized polynomial chaos bases for the stochastic dimension and eigenfunction bases for the spatial dimension. Dynamic orthogonality is used to derive closed-form equations for the time evolution of mean, spatial and the stochastic fields. The resultant system of equations consists of a partial differential equation (PDE) that defines the dynamic evolution of the mean, a set of PDEs to define the time evolution of eigenfunction bases, while a set of ordinary differential equations (ODEs) define dynamics of the stochastic field. This system of dynamic evolution equations efficiently propagates the prior parametric uncertainty to the system response. The resulting bi-orthogonal expansion of the system response is used to reformulate the Bayesian inference for efficient exploration of the posterior distribution. The efficacy of the proposed method is investigated for calibration of a 2D transient diffusion simulator with an uncertain source location and diffusivity. The computational efficiency of the method is demonstrated against a Monte Carlo method and a generalized polynomial chaos approach.

The purpose of this work is to propose and characterize fractional descriptor reduced-order perfect nonlinear observers for a class of fractional descriptor discrete-time nonlinear systems. Sufficient conditions for the existence of these observers are established. The design procedure of the observers is given and demonstrated on a numerical example.

Earthquakes yield motions of massive rock layers accompanied by vibrations which travel in waves. This paper analyses the possibility of G-type wave propagation along the plane surface at the interface of two different media which is assumed to be heterogeneous and viscoelastic. The upper layer is considered to be viscoelastic and the lower half space is considered to be an initially stressed heterogeneous half space. The dispersion equation, as well as the phase and group velocities, is obtained in closed form. The dispersion equation agrees with the classical Love type wave. The effects of the nonhomogeneity of the parameters and the initial stress on the phase and group velocities are expressed by means of a graph.

The problem of robust linear and nonlinear diagnostic observer design is considered. A method is suggested to construct the observers that are disturbance decoupled or have minimal sensitivity to the disturbances. The method is based on a logic-dynamic approach which allows us to consider systems with non-differentiable nonlinearities in the state equations by methods of linear algebra.

Searching for the shortest-path in an unknown or changeable environment is a common problem in robotics and video games, in which agents need to update maps and to perform re-planning in order to complete their missions. D* Lite is a popular incremental heuristic search algorithm (i.e., it utilizes knowledge from previous searches). Its efficiency lies in the fact that it re-expands only those parts of the search-space that are relevant to registered changes and the current state of the agent. In this paper, we propose a new D* Extra Lite algorithm that is close to a regular A*, with reinitialization of the affected search-space achieved by search-tree branch cutting. The provided worst-case complexity analysis strongly suggests that D* Extra Lite’s method of reinitialization is faster than the focused approach to reinitialization used in D* Lite. In comprehensive tests on a large number of typical two-dimensional path-planning problems, D* Extra Lite was 1.08 to 1.94 times faster than the optimized version of D* Lite. Moreover, while demonstrating that it can be particularly suitable for difficult, dynamic problems, as the problem-complexity increased, D* Extra Lite’s performance further surpassed that of D*Lite. The source code of the algorithm is available on the open-source basis.

This paper presents an alternative approach to the task of control performance assessment. Various statistical measures based on Gaussian and non-Gaussian distribution functions are evaluated. The analysis starts with the review of control error histograms followed by their statistical analysis using probability distribution functions. Simulation results obtained for a control system with the generalized predictive controller algorithm are considered. The proposed approach using Cauchy and Lévy α-stable distributions shows robustness against disturbances and enables effective control loop quality evaluation. Tests of the predictive algorithm prove its ability to detect the impact of the main controller parameters, such as the model gain, the dynamics or the prediction horizon.

The design of an interval observer for estimation of unmeasured state variables with application to drinking water distribution systems is described. In particular, the design process of such an observer is considered for estimation of the water quality described by the concentration of free chlorine. The interval observer is derived to produce the robust interval bounds on the estimated water quality state variables. The stability and robustness of the interval observer are investigated under uncertainty in system dynamics, inputs, initial conditions and measurement errors. The bounds on the estimated variables are generated by solving two systems of first-order ordinary differential equations. For that reason, despite a large scale of the systems, the numerical efficiency is sufficient for the on-line monitoring of the water quality. Finally, in order to validate the performance of the observer, it is applied to the model of a real water distribution network.

In contrast to road-based traffic, the track as well as the corresponding duty cycle for railways are known beforehand, which represents a great advantage during the development of operating strategies for hybrid vehicles. Hence the benefits of hybrid vehicles regarding the fuel consumption can be exploited by means of an off-line optimisation. In this article, the fuel-optimal operating strategy is calculated for one specified track using two hybrid railway vehicles with different kinds of energy storage systems: on the one hand, a lithium-ion battery (high-energy storage) and, on the other, a double layer capacitor (high-power storage). For this purpose, control-oriented simulation models are developed for each architecture addressing the main effects contributing to the longitudinal dynamics of the power train. Based on these simulation models, the fuel-optimal operating strategy is calculated by two different approaches: Bellman’s dynamic programming, a wellknown approach in this field, and an innovative sensitivity-based optimisation.

It has always been a priority for all nations to reduce new HIV infections by implementing a comprehensive HIV prevention programme at a sufficient scale. Recently, the ‘HIV counselling & testing’ (HCT) campaign is gaining public attention, where HIV patients are identified through screening and immediately sent under a course of antiretroviral treatment (ART), neglecting the time extent they have been infected. In this article, we study a nonlinear mathematical model for the transmission dynamics of HIV/AIDS system receiving drug treatment along with effective awareness programs through media. Here, we consider two different circumstances: when treatment is only effective and when both treatment and awareness are included. The model is analyzed qualitatively using the stability theory of differential equations. The global stabilities of the equilibria under certain conditions are determined in terms of the model reproduction number. The effects of changes in some key epidemiological parameters are investigated. Projections are made to predict the long term dynamics of the disease. The epidemiological implications of such projections on public health planning and management are discussed. These studies show that the aware populations were less vulnerable to HIV infection than the unaware population.

We consider a class of two- and three-dimensional h-refined meshes generated by an adaptive finite element method. We introduce an element partition tree, which controls the execution of the multi-frontal solver algorithm over these refined grids. We propose and study algorithms with polynomial computational cost for the optimization of these element partition trees. The trees provide an ordering for the elimination of unknowns. The algorithms automatically optimize the element partition trees using extensions of dynamic programming. The construction of the trees by the dynamic programming approach is expensive. These generated trees cannot be used in practice, but rather utilized as a learning tool to propose fast heuristic algorithms. In this first part of our paper we focus on the dynamic programming approach, and draw a sketch of the heuristic algorithm. The second part will be devoted to a more detailed analysis of the heuristic algorithm extended for the case of hp-adaptive grids.

A single-server queueing system with a marked Markovian arrival process of heterogeneous customers is considered. Type-1 customers have limited preemptive priority over type-2 customers. There is an infinite buffer for type-2 customers and no buffer for type-1 customers. There is also a finite buffer (stock) for consumable additional items (semi-products, half-stocks, etc.) which arrive according to the Markovian arrival process. Service of a customer requires a fixed number of consumable additional items depending on the type of the customer. The service time has a phase-type distribution depending on the type of the customer. Customers in the buffer are impatient and may leave the system without service after an exponentially distributed amount of waiting time. Aiming to minimize the loss probability of type-1 customers and maximize throughput of the system, a threshold strategy of admission to service of type-2 customers is offered. Service of type-2 customer can start only if the server is idle and the number of consumable additional items in the stock exceeds the fixed threshold. Stationary distributions of the system states and the waiting time are computed. In the numerical example, we show some interesting effects and illustrate a possibility of application of the presented results for solution of optimization problems.

With the rapid evolution of the distributed computing world in the last few years, the amount of data created and processed has fast increased to petabytes or even exabytes scale. Such huge data sets need data-intensive computing applications and impose performance requirements to the infrastructures that support them, such as high scalability, storage, fault tolerance but also efficient scheduling algorithms. This paper focuses on providing a hybrid scheduling algorithm for many task computing that addresses big data environments with few penalties, taking into consideration the deadlines and satisfying a data dependent task model. The hybrid solution consists of several heuristics and algorithms (min-min, min-max and earliest deadline first) combined in order to provide a scheduling algorithm that matches our problem. The experimental results are conducted by simulation and prove that the proposed hybrid algorithm behaves very well in terms of meeting deadlines.

In recent years, a growing interest has been observed in research on RNA (ribonucleic acid), primarily due to the discovery of the role of RNA molecules in biological systems. They not only serve as templates in protein synthesis or as adapters in the translation process, but also influence and are involved in the regulation of gene expression. The RNA degradation process is now heavily studied as a potential source of such riboregulators. In this paper, we consider the so-called RNA partial degradation problem (RNA PDP). By solving this combinatorial problem, one can reconstruct a given RNA molecule, having as input the results of the biochemical analysis of its degradation, which possibly contain errors (false negatives or false positives). From the computational point of view the RNA PDP is strongly NP-hard. Hence, there is a need for developing algorithms that construct good suboptimal solutions. We propose a heuristic approach, in which two tabu search algorithms cooperate, in order to reconstruct an RNA molecule. Computational tests clearly demonstrate that the proposed approach fits well the biological problem and allows to achieve near-optimal results. The algorithm is freely available at http://www.cs.put.poznan.pl/arybarczyk/tabusearch.php.

The fruit fly optimization algorithm (FOA) is a global optimization algorithm inspired by the foraging behavior of a fruit fly swarm. In this study, a novel stochastic fractal model based fruit fly optimization algorithm is proposed for multiobjective optimization. A food source generating method based on a stochastic fractal with an adaptive parameter updating strategy is introduced to improve the convergence performance of the fruit fly optimization algorithm. To deal with multiobjective optimization problems, the Pareto domination concept is integrated into the selection process of fruit fly optimization and a novel multiobjective fruit fly optimization algorithm is then developed. Similarly to most of other multiobjective evolutionary algorithms (MOEAs), an external elitist archive is utilized to preserve the nondominated solutions found so far during the evolution, and a normalized nearest neighbor distance based density estimation strategy is adopted to keep the diversity of the external elitist archive. Eighteen benchmarks are used to test the performance of the stochastic fractal based multiobjective fruit fly optimization algorithm (SFMOFOA). Numerical results show that the SFMOFOA is able to well converge to the Pareto fronts of the test benchmarks with good distributions. Compared with four state-of-the-art methods, namely, the non-dominated sorting generic algorithm (NSGA-II), the strength Pareto evolutionary algorithm (SPEA2), multi-objective particle swarm optimization (MOPSO), and multiobjective self-adaptive differential evolution (MOSADE), the proposed SFMOFOA has better or competitive multiobjective optimization performance.

In recent years, the counterparty credit risk measure, namely the default risk in over-the-counter (OTC) derivatives contracts, has received great attention by banking regulators, specifically within the frameworks of Basel II and Basel III. More explicitly, to obtain the related risk figures, one is first obliged to compute intermediate output functionals related to the mark-to-market position at a given time no exceeding a positive and finite time horizon. The latter implies an enormous amount of computational effort is needed, with related highly time consuming procedures to be carried out, turning out into significant costs. To overcome the latter issue, we propose a smart exploitation of the properties of the (local) time spent by the Brownian motion close to a given value.