Swarm Intelligence Approaches for Grid Load Balancing

With the rapid growth of data and computational needs, distributed systems and computational Grids are gaining more and more attention. The huge amount of computations a Grid can fulfill in a specific amount of time cannot be performed by the best supercomputers. However, Grid performance can still be improved by making sure all the resources available in the Grid are utilized optimally using a good load balancing algorithm. This research proposes two new distributed swarm intelligence inspired load balancing algorithms. One algorithm is based on ant colony optimization and the other algorithm is based on particle swarm optimization. A simulation of the proposed approaches using a Grid simulation toolkit (GridSim) is conducted. The performance of the algorithms are evaluated using performance criteria such as makespan and load balancing level. A comparison of our proposed approaches with a classical approach called State Broadcast Algorithm and two random approaches is provided. Experimental results show the proposed algorithms perform very well in a Grid environment. Especially the application of particle swarm optimization, can yield better performance results in many scenarios than the ant colony approach.     

A conceptual theory of part-whole relations and its applications

This paper shows two applications of a theory of part-whole relations that has been introduced in [1] in two different areas: natural language semantics and modelling parts of physical objects. A short overview of the theory will be given, followed by two sections which present results in the two application areas. In Section 2 we provide an analysis of the domain of solid objects like devices, buildings and other artifacts of our daily life. We first examine in more detail the relevant part-whole relations in that domain. The second part describes how these relations can be represented and manipulated in a hybrid system. In Section 3 it will be shown that there is a strong and immediate correspondence between classes of part-whole-relations and syntactically or lexically motivated classes of genitive constructions. The two applications illustrate the generality of the theory as well as its usefulness for representing and reasoning about part-whole information in totally different domains.     

External and internal electrostatic potentials of cholinesterase models

The electrostatic potentials for the three-dimensional structures of cholinesterases from various species were calculated, using the Delphi algorithm, on the basis of the Poisson–Boltzmann equation. We used structures for Torpedo californica and mouse acetylcholinesterase, and built homology models of the human, Bungarus fasciatus, and Drosophila melanogaster acetylcholinesterases and human butyrylcholinesterase. All these structures reveal a negative external surface potential, in the area around the entrance to the active-site gorge, that becomes more negative as the rim of the gorge is approached. Moreover, in all cases, the potential becomes increasingly more negative along the central axis running down the gorge, and is largest at the base of the gorge, near the active site. Ten key acidic residues conserved in the sequence alignments of AChE from various species, both in the surface area near the entrance of the active-site gorge and at its base, appear to be primarily responsible for these potentials. The potentials are highly correlated among the structures examined, down to sequence identities as low as 35%. This indicates that they are a conserved property of the cholinesterase family, could serve to attract the positively charged substrate into and down the gorge to the active site, and may play other roles important for cholinesterase function.     

Fracture behaviour of single crystal silicon microstructures

The fracture behaviour of single crystal silicon (SCSi) microstructures is analysed based on microme-chanical torsional and tensile experiments. The uniaxial testpieces are characterised by the presence of sharp not-ches at the gauge length extremities. The critical loading conditions are reproduced in a finite element model in order to identify the analogies of the failure conditions in tension and torsion. The stress field in the vicinity of the notch tip (were cracks originate) is analyzed, and fracture mechanics parameters are determined. In the finite element model a crack, reproducing the failure process observed in the experiments, is included. The crack area is incrementally increased and the energy release rate for the critical loading conditions in tension and torsion is calculated. Based on these results a failure criterion is formulated along with a procedure for the mechanical integrity analysis of SCSi microstructures of arbitrary shape and loading conditions.     

High-Energy Physics on the Grid: the ATLAS and CMS Experience

In this paper we present the experience of the ATLAS and CMS High-Energy Physics (HEP) experiments at the Large Hadron Collider (LHC) with the LCG/EGEE Grid infrastructure. The activity developed around the following two main lines: large-scale physics and detector simulations and end-user analysis. The LCG/EGEE Grid infrastructure offers a large amount of computing and storage resources and is growing very rapidly. It provides the natural environment for large-scale physics and detector simulations. Also, the analysis of these detector simulation data (and in the near future of the reconstructed data from physics collisions) requires efficient end-users access to Grid resources. In this paper, the main findings and lessons learned in terms of performance, robustness and scalability of the whole system are discussed in detail.     

Performance evaluation of semantic registries: OWLJessKB and instanceStore

Service discovery is a critical task in distributed computing architectures for finding a particular service instance. Semantic annotations of services help to enrich the service discovery process. Semantic registries are an important component for the discovery of services and they allow for semantic interoperability through ontology-based query formulation and dynamic mapping of terminologies between system domains. This paper evaluates two semantic registries—OWLJessKB implementation and instanceStore—to determine the suitability of these with regards to the query response time and the overall scalability for use in mathematical services. Mathematical ontologies from the MONET project are used to undertake comparison. The results demonstrate that the performance of registries may be compared across two axes: (1) time to initialize (i.e. time to load an initial ontology into memory); (2) time to query (i.e. time to reason with an ontology loaded into memory).     

Modeling the impact of spectral sensor configurations on the FLD retrieval accuracy of sun-induced chlorophyll fluorescence

Chlorophyll fluorescence is related to photosynthesis and can serve as a remote sensing proxy for estimating photosynthetic energy conversion and carbon uptake. Recent advances in sensor technology allow remote measurements of the sun-induced chlorophyll fluorescence signal (Fs) at leaf and canopy scale. The commonly used Fraunhofer Line Depth (FLD) principle exploits spectrally narrow atmospheric oxygen absorption bands and relates Fs to the difference of the absorption feature depth of a fluorescensing and a non-fluorescensing surface. However, due to the nature of these narrow bands, Fs retrieval results depend not only on vegetation species type or environmental conditions, but also on instrument technology and processing algorithms. Thus, an evaluation of all influencing factors and their separate quantification is required to further improve Fs retrieval and to allow a reproducible interpretation of Fs signals.Here we present a modeling study that isolates and quantifies the impacts of sensor characteristics, such as spectral sampling interval (SSI), spectral resolution (SR), signal to noise ratio (SNR), and spectral shift (SS) on the accuracy of Fs measurements in the oxygen A band centered at 760 nm (O₂-A). Modeled high resolution radiance spectra associated with known Fs were spectrally resampled, taking into consideration the various sensor properties. Fs was retrieved using the three most common FLD retrieval methods, namely the original FLD method (sFLD), the modified FLD (3FLD) and the improved FLD (iFLD). The analysis investigates parameter ranges, which are representative for field and airborne instruments currently used in Fs research (e.g., ASD FieldSpec, OceanOptics HR, AirFLEX, AISA, APEX, CASI, and MERIS).Our results show that the most important parameter affecting the retrieval accuracy is SNR, SR accounts for ≤ 40% of the error, the SSI for ≤ 12%, and SS for ≤ 7% of the error. A trade-off study revealed that high SR can partly compensate for low SNR. There is a strong interrelation between all parameters and the impact of specific parameters can compensate or amplify the influence of others. Hence, the combination of all parameters must be considered by the evaluation of sensors and their potential for Fs retrieval. In general, the standard FLD method strongly overestimates Fs, while 3FLD and iFLD provide a more accurate estimation of Fs. We conclude that technical sensor specifications and the retrieval methods cause a significant variability in retrieved Fs signals. Results are intended to be one relevant component of the total uncertainty budget of Fs retrieval and have to be considered in the interpretation of retrieved Fs signals.     

Artificial neural networks for document analysis and recognition

Artificial neural networks have been extensively applied to document analysis and recognition. Most efforts have been devoted to the recognition of isolated handwritten and printed characters with widely recognized successful results. However, many other document processing tasks, like preprocessing, layout analysis, character segmentation, word recognition, and signature verification, have been effectively faced with very promising results. This paper surveys the most significant problems in the area of offline document image processing, where connectionist-based approaches have been applied. Similarities and differences between approaches belonging to different categories are discussed. A particular emphasis is given on the crucial role of prior knowledge for the conception of both appropriate architectures and learning algorithms. Finally, the paper provides a critical analysts on the reviewed approaches and depicts the most promising research guidelines in the field. In particular, a second generation of connectionist-based models are foreseen which are based on appropriate graphical representations of the learning environment.     

When simulated environments make the difference: the effectiveness of different types of training of car service procedures

An empirical analysis was performed to compare the effectiveness of different approaches to training a set of procedural skills to a sample of novice trainees. Sixty-five participants were randomly assigned to one of the following three training groups: (1) learning-by-doing in a 3D desktop virtual environment, (2) learning-by-observing a video (show-and-tell) explanation of the procedures, and (3) trial-and-error. In each group, participants were trained on two car service procedures. Participants were recalled to perform a procedure either 2 or 4 weeks after the training. The results showed that: (1) participants trained through the virtual approach of learning-by-doing performed both procedures significantly better (i.e. p < .05 in terms of errors and time) than people of non-virtual groups, (2) the virtual training group, after a period of non-use, were more effective than non-virtual training (i.e. p < .05) in their ability to recover their skills, (3) after a (simulated) long period from the training—i.e. up to 12 weeks—people who experienced 3D environments consistently performed better than people who received other kinds of training. The results also suggested that independently from the training group, trainees’ visuospatial abilities were a predictor of performance, at least for the complex service procedure, adj R ² = .460, and that post-training performances of people trained through virtual learning-by-doing are not affected by learning styles. Finally, a strong relationship (p < .001, R ² = .441) was identified between usability and trust in the use of the virtual training tool—i.e. the more the system was perceived as usable, the more it was perceived as trustable to acquire the competences.