Evaluating fault tolerance approaches in multi-agent systems

A multi-agent system (MAS) is a distributed system that consists of multiple agents working together to solve mutual problems. Even though MASs are well suited for the development of complex distributed systems, the number of real-world usages is still small. One of the main reasons for this is that MASs are very fragile. In a typical, large-scale MAS, the rate of failure grows with the number of hosts, the number of deployed agents, and the duration of the agent’s task execution. For this reason, numerous approaches have been introduced to deal with aspects of failure handling. However, the absence of centralized control and a large number of individual intelligent components makes it difficult to detect and treat errors. The risk of uncontrollable fault propagation is high and can seriously impact on system performance. There are two important factors that limit the usage of MASs: (1) existing fault tolerance (FT) approaches are not generic, as they focus on and improve specific issues of FT; and (2) despite the plethora of available FT approaches and theories, there is a remarkable lack of general metrics, tools, benchmarks, and experimental methods for formal validation and comparison of existing or newly developed FT approaches. As FT approaches in MASs become a well-established field, the need for generalized, standardized evaluation of FT approaches emerges as imperative. In this paper, we first present a detailed overview of existing FT solutions, approaches, and techniques in agent platform hosted MASs. From that overview, we derive the commonalities in existing research. Next, we present the main contribution of our paper: an evaluation methodology, with a set of metrics, for comparing FT approaches in MASs. We adopt an engineering perspective on the problem, defining a methodology and metrics that are both implementation- and domain-independent. The metrics are formalized with an acyclic directed graph. By using our methodology, evaluators can select an appropriate FT approach for targeted MAS application, thus improving MAS usability, stability, and development speed. In order to show the viability of our approach, a case study that compares two FT approaches for a targeted MAS is presented. The case study results show that our methodology can be used for selecting an appropriate FT approach for the targeted MAS.     

Cycling of electrodeposited beta lead dioxide

The electrochemical behaviour of lead dioxide cycled in sulphuric acid under galvanostatic (discharge) and potentiostatic (charge) conditions has been investigated. The results illustrate the effect of the sulphate barrier layer forming on PbO₂. This layer shifts the equilibrium potential of the electrodes without influence on the kinetics of the reduction, which seems to be dependent on a diffusive step associated with the onset of an O^2– gradient in the dioxide. As for the sulphate oxidation, it has been found that this process is mainly affected by the end potential reached during the previous galvanostatic reduction.     

PtRu Colloid Nanoparticles for CO Oxidation in Microfabricated Reactors

The catalytic activity of PtRu colloid nanoparticles for CO oxidation is investigated in microfabricated reactors. The measured catalytic performance describes a volcano curve as a function of the Pt/Ru ratio. The apparent activation energies for the different alloy catalysts are between 21 and 117 kJ/mol, which agree well with literature. The size distribution of the colloid nanoparticles is determined from STM and TEM, from which an average size of the colloid nanoparticles of 2.2±0.5 nm is determined.     

Changes in the amount of kidney pigmented macrophage aggregates throughout the breeding cycle of female Ohrid trout, Salmo letnica Kar. (Teleostei, Salmonidae)

Changes in fish macrophages (Macs) are useful indicators of environmental pressures, but responses due to chemical and nonchemical stresses may be confounded by natural sources of variability. These may include sex and gonadal stage. In this study, we start addressing the following question: is the seasonally dependent ovary stage a factor to consider when using kidney Macs as biomarkers? To tackle this problem, the relative amount of pigmented Macs in kidney (head, trunk, and tail portions) was stereologically estimated in Ohrid trout, and related with the breeding status. The amount of Macs significantly increased from pre vitellogenesis to late vitellogenesis and showed a decreasing trend then after, with lower values noted after spawning in the head (1.9% versus 7.5% versus 2.0%), trunk (1.8% versus 7.5% versus 2.5%), and tail (2.5% versus 6.7% versus 2.9%) kidney. The decrease seen at spawning was significant in head and trunk kidney, and at post spawning it was significant for all kidney portions. The amounts of Macs were positively correlated with the ovary relative weights and plasma estradiol levels. We proved for the first time that fish kidney pigmented Macs can vary in amount during the breeding cycle. Our data, combined with literature, strongly support that the sex‐steroid profile and kidney status‐seasonal remodeling both influence the Macs pool; likely not only in female trout. So, while increases in Macs may warn of ecosystem problems, we show that using kidney Macs for biomonitoring should also take into account seasonally, particularly that related with ovary maturation. Microsc. Res. Tech. 2011. © 2011 Wiley Periodicals, Inc.     

Assessment of powder flow characteristics in incoherent soup concentrates

Soup mixtures represent specific flow characteristics while particles of different sizes and properties form a homogeneous mixture. In such mixtures, particle–particle interactions differ with addition of different fat types. This study was done to present a characterization and comparison of the cohesion index, powder flow speed dependency and caking properties of four different aggregates of fat particles with three different moisture contents, used in various compositions of powdered soup concentrates. ESEM micrographs have shown that different fat types bind particles differently. The moisture content of cream soup concentrates has a significant influence on cake height ratio of all cycles – more moisture causes higher cake height ratios (from r_s = 0.86 to r_s = 0.76; p < 0.05). Cream soup concentrates cake strength values are also significantly influenced by the mixtures moisture content – higher moisture content samples showed higher cake strength values (r_s = 0.73, p < 0.05). There is a significant correlation between compaction coefficients of all speeds and cake height ratios of all cycles for the cream soup concentrates. Combination of measuring techniques (Powder Flow Analyzer, ESEM and Mastersizer) together with Spearman’s Rank Order Correlation, as a method of non-parametric statistics, provides parturient results in characterization of extremely non-homogenous powder mixtures.     

High-throughput SEM preparation of proteinaceaous extracellular matrix

Adhesives of aquatic organisms are of high scientific interest with a view to biomimicry. The visualization of their fine-structure, however, is difficult due to the fragility of the extracellular matrix (ECM). Care must be taken in its preparation for high-resolution SEM. Relating matrix structures to substrate properties demands a high throughput of samples for reliable comparisons. In order to acquaint ourselves with a suitable method we found an easy way to manage the critical steps of preparation for SEM. We show that thin layers of proteinaceous matrix can be satisfactorily prepared by combining freeze-drying and sputtering in a conventional sputter coater after one-step fixation (2.5% glutaraldehyde). The results were superior to CPD.     

An impact of Croatian journals measured by citation analysis from SCI-expanded database in time span 1975–2001

The aim of this research is to gain an insight into international recognition of the STM (Science, Technology, and Medicine) Croatian journals measured by citations in SCI-expanded database. The sample for the research was a citation analysis of 142 journals in time span 1975–2001 for papers published in 1975–1998. More than 90% of those journals are not indexed by SCI-expanded. For the purpose of this research we introduced a new scientometric indicator Normalized number of Citations per 100 Papers (NCP) that allows us direct comparison of the journals from various categories (NCP = 100C/P / IF₁₉₈₉). We chose the year 1989 as a mean value for time span 1975–2001. By citation analysis we established the influence of errors on recognition of Croatian journals and their articles. Obtained results show that an article-to-article link is not found for 32% of cited items. The most frequent type of error is journal title, 37%, which indicates that approximately one third of Croatian journals can not be found when searching by journal title only. Some Croatian journals, even not indexed by SCI-expanded, showed relatively high rank in an impact, i.e. their NCP is higher than 100, and number of citations per paper is higher than 1.     

Face Mask Extraction in Video Sequence

International Journal of Computer Vision - Inspired by the recent development of deep network-based methods in semantic image segmentation, we introduce an end-to-end trainable model for face mask...