Computational Science – ICCS 2006

The term "the Grid" was coined in the mid 1990s to denote a proposed distributed computing infrastructure for advanced science and engineering. Grid computing technology is a new way for remotely sensed data processing. Special remotely sensed data processing Grid node provides convenient, consistent spatial information processing service for user. In this paper we mainly discusses information registry mechanism and method for remotely sensed data processing module and algorithm in Grid environment and gives the remotely sensed meta-module conception. Then based on the characteristic of remotely sensed data processing module and Grid technology, we describe remotely sensed data processing meta-module information registry method and give a concrete example.

The outbreak of Avian Influenza A (H5N1) infection has spread across all over the world from East-South Asia to Russia, Greece, Romania and Turkey. It will be important to find the transmission route and determine the environmental factor that affect the prevalence of avian influenza A virus. Based on the environmental parameters derived from remote sensing (RS) measurements and the avian influenza A (H5N1) infection case data in China during January 23, 2004 to February 24, 2004, the correlations between the outbreak of H5N1 avian influenza and the environmental parameters of the infected area, such as land surface temperature, was conducted using the spatial analysis abilities of GIS. The statistically significant association between the land use or land cover and outbreak of avian influenza A infection was found, i.e. about 86.4% of the 44 cases are in the cropland. Besides, by the buffering analysis, it is estimated that the vicinity at 50 km or so to main railways plays a key role in the spatial distribution of avian influenza A infection. Finally, we draw preliminary conclusion that the infection often outbreak in a certain range of land surface temperature etc probably due to in part the H5N1 virus implications.

In this paper an alternative non-pointer quadtree node codification to manage geographical spatial data is presented. New codification is based on a variable sequence of z-ordered base four digits. Memory requirements of the new codification are lower than previous codifications, and in particular lower than FD codification, the most commonly used in linear quadtrees. Furthermore, z-ordering makes compatible new codification with most of the algorithms developed for FD.

This paper mainly discusses the special problems and solutions for multi-task and data flow control in cluster parallel computing system which dedicated used for High-performance Remote Sensing Satellite Ground Pre-processing System (GHIPS). After giving the overview of the GHIPS, the structure and function of Operation and Mission Subsystem (OMS) shall be formulated. The more detail discussion shall be focused on the organization and processing mechanism based on workflow of the task procedure as well as the task scheduling strategies. Based on our experiences more flexible and reasonable solutions will be given.

Hyperspectral imaging is a new technique in remote sensing that generates hundreds of images corresponding to different wavelength channels for the same area on the surface of the Earth. Most available techniques for hyperspectral image classification focus on analyzing the data without incorporating the spatial information; i.e. the data is treated not as an image but as an unordered listing of spectral measurements where the spatial coordinates can be shuffled arbitrarily without affecting the final analysis. Despite the growing interest in the development of techniques for interpretation and classification of such high-dimensional imagery, only a few efforts devoted to the design of parallel implementations exist in the open literature. In this paper, we describe AMEEPAR, a parallel morphological algorithm that integrates the spatial and spectral information. The algorithm has been specifically optimized in this work for execution on heterogeneous networks of workstations. The parallel properties and classification accuracy of the proposed approach are evaluated using four networks of workstations distributed among different locations, and a massively parallel Beowulf cluster at NASA’s Goddard Space Flight Center.

The development of new tools and methodologies is necessary in order to better understand current and past climatic changes,. To be useful, these mathematical or software tools must not remain only in the hands of specialists in statistics, but must also be usable by the larger community of paleoclimatologists. It is therefore necessary to conceive a user interface adapted to the specificities of their use in paleoclimatology. Here, we propose the development of new tools of interactive analysis. Through the combination of techniques coming from knowledge discovery and information visualization (visual data mining), rapid and accurate paleoclimatic reconstructions will be easier to produce.

This paper presents a novel approach to terrain rendering, which mostly relies on GPU/shader rather than CPU. The most popular representation for terrain data is uniformly sampled height field. As the height field is stored as a texture map, it is directly accessible by a pixel shader. The pixel shader uses a ray casting algorithm, and the CPU and the vertex shader provide ray information to be passed to the pixel shader. Then, the pixel shader samples the ray, computes the intersection of the ray and the terrain surface, and finally determines the pixel color. The experimental results show the feasibility of the shader-intensive approach to real-time terrain rendering.

Spherical Harmonic Transforms (SHTs) which are essentially Fourier transforms on the sphere are critical in global geopotential and related applications. For analysis purposes, discrete SHTs are difficult to formulate for an optimal discretization of the sphere, especially for applications with requirements in terms of near-isometric grids and special considerations in the polar regions. With the enormous global datasets becoming available from satellite systems, very high degrees and orders are required and the implied computational efforts are very challenging. Among the best known strategies for discrete SHTs are quadratures and least squares. The computational aspects of SHTs and their inverses using both quadrature and least-squares estimation methods are discussed with special emphasis on information conservation and numerical stability. Parallel and grid computations are imperative for a number of geodetic, geophysical and related applications, and these are currently under investigation.

The aim of the present paper is to propose a method of conditional stochastic simulation of propagation of seismic wave using the spatiotemporal correlation function. The method has been used to generate unknown time histories at various points of ground motion random filed based on the specified earthquake record at one location. The results of the study show that the method considered gives relatively low simulation errors.

This paper presents the development of a virtual reality urban planning tool based on GIS and Virtual Environments to facilitate scenario generation and collaborative urban design. Our system provides an open structure for users to support various visualisation and simulation modules, for the prototyping of urban designs. An Urban Scene Generation module is used within this urban planning system for generating the urban layout, on which the various collaborative planning tasks are carried out. A 3D urban model is generated from a GIS database by extracting features such as landscape, roads, and buildings. The system allows the users to place CAD models within the virtual environment. Various statistical data related to population, crime, and employment are also visualised in this virtual environment. A VR based user interface is provided to handle urban simulation data, query information and discuss urban simulation scenarios.

We present ongoing research in the Resurgence (RESearch sURGe ENabled by CyberinfrastructurE) project. This infrastructure shall enable the flexible combination of computational chemistry tools from a unified interface, with the focus on automated high-throughput processing. The implementation is based on the idea that the time-consuming parts of the calculations can be distributed onto computational Grids using the Kepler scientific workflow system and the Nimrod toolkit for distributed parametric modeling. We describe an example workflow that allows preparing, running, and displaying jobs on different molecules, employing the GAMESS quantum chemical program package.

This study is based on our earlier work with reactivity index to propose intra and inter molecular reactivity in catalytic materials using density functional theory, within the domain of hard soft acid base (HSAB) principle. We have as well shown a small example to show directly the utility of this method in elucidating acidity in catalytic material of interest. Our goal is to show that a simple theory can be useful to design new futuristic material of interest.

Binding energy spectra and orbital momentum distributions of the two most stable conformers of L-alanine are investigated. Molecular properties such as geometry and dipole moments agree well with available experimental and previous theoretical investigations. Dual space analysis is employed to study the binding energy spectra in coordinate space based on B3LYP/TZVP density functional calculations, and the valence orbital momentum distributions based on the plane wave impulse approximation. In the valence space, the HOMO (24

a

), NHOMO (23

a

) and orbitals 22

a

and 18

a

are selected to study the conformational processes in L-alanine.

Electronic excitations are key points of most of the commonly measured optical spectra. The first principle studies of excited states however require much larger effort than computations of the ground state reliably reproduced by the density functional theory (DFT). In present work computation of optical functions of organic molecular complexes is studied. The system of independent particles excited by external light field is considered within perturbation theory (the random phase approximation, RPA). Optical response functions are calculated using

ab initio

pseudopotentials theory. Results of predicted optical absorption associated with organic semi-conducting conjugated polymers, poly-phenylene-vinylenes (PPV), are presented. Effects of different corrections to the DFT improving accuracy are considered. Results are discussed in comparison with available experimental data.

Grid computing is a promising technology for computational chemistry, due to the large volume of calculations involved in appplications such as molecular modeling, thermochemistry and other types of systematic studies. Difficulties in using computational chemistry codes in grid environments arise, however, from the fact that the application software is complex, requiring substantial effort to be installed on different platforms. Morever, these codes depend upon task–dependent sets of data files to be present at the execution nodes. Aiming to improve the usability of different quantum chemistry codes in the distributed, heterogeneous environments found in computational grids, we describe a framework capable of handling the execution of different codes on different platforms. This framework can be divided into three independent parts, one dealing with the mapping of a calculation to a set of codes and the construction of execution environments, one dealing with the management of grid resources, and one that takes care of the heterogeneity of the environment. The suitability of this framework to tackle typical quantum chemistry calculations is discussed and illustrated by a model application.

Numerical prediction of the physical properties of nanocomposites is an attractive area that requires more discussion and investigation. In this paper, we calculate the thermal conductivity of nanocomposites embedded with carbon nanotubes (CNTs) based on the representative volume element (RVE) concept. The RVE, which encompasses a single CNT, was constructed assuming that the CNTs are distributed in polymeric material homogeneously, and also assuming that the CNTs have no interaction with other CNTs. This research describes the thermal characteristics of nanoscale materials – CNTs filled nanocomposites – as a case study and measured their thermal conductivity, for the purpose of validation of numerical results. The dispersion state of the CNTs was observed using field emission scanning electronic microscope (FESEM). We found that the numerically predicted thermal conductivity is closely matches the experimental one and that the numerical tool employed in the study is superior to other analytical and numerical methods.

The current work focuses on resin bleeding process during the Seeman composite resin infusion molding process (SCRIMP), which is a subset of liquid composite molding (LCM) process. Finite difference method (FDM) is implemented to predict the preform thickness, bleeding resin volume, and fiber volume fraction by using a non-rigid control volume. After the fibrous preform is completely impregnated, the resin flow within the preform has a great impact on the dimension and mechanical properties of the final composite parts. As the resin flows out of the preform, the resin pressure and preform thickness are reduced, which increases the fiber volume fraction and the dimension tolerance of the preform.

In this paper, the influence of resin flow rate at vent in the mold is also investigated. It is found that there is a critical flow rate to optimize the SCRIMP process at the vacuum line.

This paper presents the new stiff solvers of the new version 2.2 of the Kinetic PreProcessor (KPP). Taking a set of chemical reactions and their rate coefficients as input, KPP generates Fortran90, Fortran77, Matlab, or C code for the temporal integration of the kinetic system. Efficiency is obtained by carefully exploiting the sparsity structures of the Jacobian and of the Hessian. A set of integration methods was added to the comprehensive suite of stiff numerical integrators. Moreover, KPP is now ready do be used to generate the tangent linear model, as well as the continuous and discrete adjoint models of the chemical system to do sensitivity analysis.

Electronic and geometrical structures of Fe

4

C

n

(CO)

m

(n+m≤6) and Fe

4

C

n

(n=7—16) clusters along with their singly negatively and positively charged ions are computed using density functional theory with generalized gradient approximation (DFT-GGA). Isomers with CO bonded directly to the iron atoms and bonded to a carbon atom chemisorbed on the cluster surface are optimized for the Fe

4

C

2

CO, Fe

4

C

2

(CO)

2

, Fe

4

C

3

CO, and Fe

4

C

4

CO series. The computed total energies are used to estimate the energetics of the Boudouard disproportionation reactions Fe

4

C

n

(CO)

m

+ CO ( Fe

4

C

n + 1

(CO)

m − − 1

+ CO

2

. Optimizations of the Fe

4

C

4

–Fe

4

C

16

clusters have shown that dimers C

2

are formed in the lowest energy states of Fe

4

C

4,

, trimers C

3

– in Fe

4

C

5

and Fe

4

C

6,

tetramers C

4

–in Fe

4

C

7

and Fe

4

C

8

, a pentamer C

5

– in Fe

4

C

9

, and a hexamer C

6

– in Fe

4

C

10

. C

n

rings attached to a Fe

3

face are formed in the lowest energy states of Fe

4

C

n

beginning with n=11.

Separation of stereoisomers of organic compounds is an important and challenge task for chemists. Cyclodextrins and their derivatives have been widely used in chromatography for this application. Experimental results indicated that substituents on the hydroxyl groups of cyclodextrin affect the efficiency of the chiral separation of

β

-butyrolactone. The understanding of the interactions contributed to the chiral recognition of cyclodextrin would help us predict the separation capability of a specific pair of cyclodextrin and chiral compound. Thus, the cyclodextrin substituent effect on the chiral recognition should be systematically investigated. In this study, Hartree Fock method with 3-21G basis set and density functional theory B3LYP with 6-31G* basis set were applied to determine the chiral recognition of a chiral model,

β

- butyrolactone, by

β

-cyclodextrin and its derivatives. Both methods predicted comparable values of chiral recognition of

β

-cyclodextrin derivatives. We found that methoxyl substitution on the wider rim of cyclodextrin (secondary hydroxyl groups) give the most effective chiral separation (ΔΔE=18.2 kcal/mol in favor of R-isomer) followed by substitution on the narrow rim (ΔΔE=9.5 kcal/mol in favor of S-isomer) while substitution on both side give the worst recognition (ΔΔE=3.2 kcal/mol in favor of S-isomer). This suggests that

β

-cyclodextrin with substitution only on the wider rim give the best chiral selectivity. By replacing methyl group with chiral hydroxypropyl group, we found that the chiral selectivity is reduced (ΔΔE=6.4 and 8.4 kcal/mol respectively for R- and S-form of hydroxypropyl group). This implies that the bulky group causes the reduction of the chiral selectivity.

The insertion reactions of singlet mono and dichlorocarbenes (

1

CHCl and

1

CCl

2

) into primary, secondary and tertiary C-H bonds of methane, ethane, propane,

n

-butane and

iso

-butane have been investigated at ab initio (MP2) and DFT levels invoking 6-31g(d, p) basis set. Among the

σ

and

π

insertions into the said alkane C-H, both MP2 and DFT predict that the

σ

approach is slightly favoured over the

π

approach. For inverted

σ

approach the barrier heights have been found to be

ca

. 2 to 5 times that of the normal

σ

approach.

This is a systematic theoretical gas phase study of the gauche and trans conformers of 1-fluoro-2-haloethanes (FCH

2

CH

2

X, X=Cl, Br and I). The methods used are second order Møller-Plesset theory (MP2) and density functional theory (DFT). The basis set used is 6-311++(d,p) for all atoms except that 3-21G and CRENBL ECP are used for iodine atom. The functional used for DFT method is B3LYP. G2/MP2 calculation has also been carried out using MP2 optimised structure. The results indicate that unlike 1,2-difluoroethane, there is the absence of gauche effect and thus there is more preference for the trans conformer. The preference for the more stable trans conformer increases with increasing atomic size of the substituted halogen atom. The same trend is observed for energy difference between the gauche and trans conformers. The 1-fluoro-2-haloethanes have also been subjected to vibrational analysis.

Constrained molecular dynamics simulations have been used to investigate ion pairing in water – DMSO mixtures. The potentials of mean force between the sodium – chloride ion pair are constructed by estimating the mean forces between the ion pair at various interionic separations and then integrating the mean forces. Two compositions of the solvent mixture with DMSO mole fractions of 0.21 and 0.35 are considered. Two model potentials for water and DMSO have been considered. One of the main observations is that the contact ion pair which is dominantly present in both the individual solvents is conspicuously absent in the mixture compositions studied. While solvent separated ion pairs dominate in all the mixture compositions, there is a presence of a second solvent separated ion pair in the water-DMSO mixture of composition with mole fraction of DMSO = 0.21. The potentials of mean force are verified by dynamical trajectories of the ion pair. The dynamics of the solvation shells has also been investigated in detail.

The subject of this paper is the theory of fault distinguishable discrete event systems. Any such system is modelled by a live, bounded, and reversible place-transition net. The notions of D-partition of the set of places P of a given place-transition net N and net k-distinguishability are first introduced. The system k-distinguishability measure is obtained in a unique way from the place-invariant matrix. For a large value of k, the system model is extended by using some set of additional places called test points. It is shown that the test point placement process will not change the above-assumed original net properties. Several examples are given.

An alternative approach to modelling and analysis of interactions among agents in multiagent systems (MAS) and their control is presented in analytical terms. The reachability graph of the Petri net (PN)-based model of MAS is found as well as the space of feasible states. Trajectories representing the interaction processes among agents in MAS are computed by means of the mutual intersection of both the straight-lined reachability tree (from a given initial state towards the terminal one) and the backtracking reachability tree (from the desired terminal state towards the initial one however, oriented towards the terminal state). Control interferences are obtained on base of the most suitable trajectory chosen from the set of feasible ones.

With the development of wireless communication technology, mobile business become more and more popular. Using GPRS or WAP protocals, the wireless devices can connect to the Web servers, retrieve information from the online databases and run special application programs. Because of the limitation of the wireless communication and mobile computing enviroment, it is difficult to improve the execution efficiency for the program that located in mobile devices. To solve the problem, introducing the cache mechanism is the major and effective method. But the traditional cache model can not achieve an acceptable cache hit ratio. The semantic caching is particularlly attractive in a mobile business environment, due to its content-based reasoning ability and semantic locality. In semantic-driven cache model, only the required data is transmitted to wireless device. In this paper we propose an application-oriented semantic cache model. It establishes an semantic associated rule-base according to the knowledge of application domains, makes use of the semantic locality for data prefetching, and adopts a Two-level LRU algorithm for cache replacement. Several experiments demonstrate that the semantic-driven cache model can achieve higher hit ratio than traditional models.

Event based system development is increasingly becoming popular for large-scale and heterogeneous distributed platforms because it helps diminishing software dependencies, and enhancing system integration and evolution. The architecture of an event based system should be tolerant to error and network fallout especially in dispatching service. Throughout the entire design of event based systems, fault-tolerance mechanism plays very important role in developing large scale middleware. This is a crucial quality of service where node failures are frequent in wide area networks with many brokers. In this paper, we address fault tolerance mechanism of the agent based distributed event system where events are responsible for determining their own paths, in the case of link and broker failures. This mechanism is achieved by dynamically configuring new paths at run time for making the system more scalable and robust on a global scale.

The primary issues in the development of advanced traveler information systems (ATIS) within the intelligent transportation systems (ITS) framework are the optimal estimation of freeway travel time and incident detection with reasonable accuracy. Typically ATIS aims to provide route guidance based on the traveler’s requirements using the information gathered from various sources such as loop detectors and probe vehicles. Until recent times traffic information was collected from mostly stationary devices and analyzed. In this research paper we consider data acquired form primarily GPS-based sources. The aim of the research is a comprehensive analysis of collected information from GPS sources using the fuzzy c-means algorithm (FCM) which provides the estimation of link speed. The modified FCM is used to extract patterns from the traffic data collected from a busy network of downtown streets. The link speed estimation is performed using smoothing techniques. Finally we apply the neuro-fuzzy algorithm to the task of incident detection from the traffic patterns.

This paper presents a consensus-based approach utilized within a multi-agent system which assists users in retrieving information from the Internet. In this system consensus methods are applied for reconciling inconsistencies among independent answers generated by agents (using different search engines) for a given query. Proposed agent system has been implemented and initial experimental results are presented.

In recent years, kNN algorithm is paid attention by many researchers and is proved one of the best text categorization algorithms. Text categorization is according to training set which is assigned class label to decide a new document which is not assigned class label belongs to some kind of document. Until now, kNN algorithm has still some issues to need to study further. Such as: improvement of decision rule; selection of k value; selection of dimensions (i.e. feature set selection); problems of multiclass text categorization; the algorithm’s executive efficiency (time and space) etc. In this paper, we mainly focus on improvement of decision rule and dimension selection. We design an adaptive fuzzy kNN text classifier. Here the adaptive indicate the adaptive of dimension selection. The experiment results show that our algorithm is effective and feasible.

The aim of this paper is to propose an architecture of distributed Intrusion Detection System (IDS). It is assumed that IDS system will detect and track dissemination and activity of the Internet worms. A general architecture for such a distributed multiagent system is proposed and the tasks, techniques and algorithms to be used are sketched.

The general aim of this paper is to study the spatio-temporal modeling techniques which can efficiently represent multiple moving objects’ in video databases. The traditional schemes only consider direction property, time interval property, and spatial relationship property for modeling moving objects’ trajectories. But, our scheme also takes into account on distance property, conceptual location information, and related object information so that we may improve a retrieval accuracy to measure a similarity between two moving objects as well as them. As its application, we implement the Content- and Semantic-based Soccer Video Retrieval (CS

2

VR) system by using MS Visual C++ and DirectX for indexing and searching on soccer video data. The CS

2

VR helps users to easily extract the trajectory information of soccer ball form soccer video data semi-automatically as well as to conveniently retrieve the results acquired by sketching query trajectory with mouse button.

Most of the network service specifications use XML based models. However, as XML imposes a hierarchical structure, several types of relations may not be modeled. Therefore, richer specification languages are required in order to specify all network services vocabulary and how it relates with management tasks and with network configuration. This paper presents an ontology based model for network services, overcoming those semantic gaps and creating a better ground for reasoning over services fostering their self-configuration.

In this paper, we propose a novel synchronization method based on contextual information elicited from a group of peers for online collaborative browsing on p2p environment. Thereby, the users are semantically tracked for modeling the context about their information searching tasks. The co-browsing system embedding our proposed method was shown to improve 52.7% and 11.5% communication performance, compared to single browsing and the asynchronous system, respectively.

In this paper, we examine pedestrian population dynamics using agent-based cellular automata models. Each pedestrian is treated as an agent, mapped onto a 2-dimensional grid. The behaviour of each agent is modelled as a sequence of specific choices reflecting different levels of autonomy. Simulations of bi-directional agent movement for four behaviours in different environments (corridors of different widths with permanent blocks such as walls) are conducted in order to identify outcomes of the behaviours and recommend a strategy. The results suggest that the “lookahead” behaviour, whilst similar to the “deterministic” behaviour, was strategically the best. Little difference was found between the “floor fields” and “random walk” behaviours.

The Nagel-Schreckenberg model of traffic is modified by the assumption that each car has an individual velocity limit. By simulations, the effect of supplementary rules is checked: (a) a speed limit of the slowest car is changed and

$\slash$

or (b) a speed limit of a car with zero gap behind is increased . It is shown that both rules increase the mean velocity; (b) rule influences the character of congested traffic – cars move though at low velocity.

In this paper is described a fast Path-Planner for Multi-robot composed by mobile robots having generic shapes and sizes (user defined) and different kinematics. We have developed an algorithm that computes the shortest collision-free path for each robot, from the starting pose to the goal pose, while considering their real shapes, avoiding the collisions with the static obstacles and the other robots. It is based on a directional (anisotropic) propagation of attracting potential values in a 4D Space-Time, using a Multilayered Cellular Automata (MCA) architecture. This algorithm searches for all the optimal collision-free trajectories following the minimum valley of a potential hypersurface embedded in a 5D Time-Space.

Modeling, designing and analyzing large scale

multi-agent systems

(MAS) with anywhere from tens of thousands to millions of autonomous agents will require mathematical and computational theories and models substantially different from those underlying the study of small- to medium-scale MAS made of only dozens, or perhaps hundreds, of agents. In this paper, we study certain aspects of the global behavior of large ensembles of simple reactive agents. We do so by analyzing the collective dynamics of several related models of discrete complex systems based on cellular automata. We survey our recent results on dynamical properties of the complex systems of interest, and discuss some useful ways forward in modeling and analysis of large-scale MAS via appropriately modified versions of the classical cellular automata.

A parallel implementation for distributed-memory MIMD systems of a 2D discrete model of laser dynamics based on cellular automata is presented.The model has been implemented on a PC cluster using a message passing library. A good performance has been obtained, allowing us to run realistic simulations of laser systems in clusters of workstations, which could not be afforded on an individual machine due to the extensive runtime and memory size needed.

In this paper we present an ongoing research that aims at providing an interpretation and detection method for spatial patterns supporting ecosystem management in the study of forest systems according to a distributed modeling and simulation approach. To this aim an innovative analysis method inspired by the Chinese Go game is under design. The originality of the approach concerns the detection within system configurations of known patterns whose interpretations are well–known by expert Go players.

The main goal this paper is to describe a software simulating spatio-temporal Dengue epidemic spread based on the utilization of a generalized probabilistic cellular automata computational analysis as the dynamic model of spatial epidemiology. This epidemic spatial model permits to reproduce explicitly the interaction of two types of transmission mechanisms in terms of global and local variables, which in turn can be adjusted to simulate respectively the populational mobility and geographical neighborhood contacts. The resulting virtual laboratory was designed to run spatio-temporal simulation of the Dengue disease spreading based on local and global interactions among two distinct populations (humans and mosquitoes).

Complex stochastic Boolean systems, depending on a large number

n

of statistically independent random Boolean variables, appear in many different scientific, technical or social areas. Each one of the 2

n

binary states associated to such systems is denoted by its corresponding binary

n

-tuple of 0s and 1s,

$\left( u_{1},\ldots,u_{n}\right) $

, and it has a certain occurrence probability

$\Pr\left\{ \left( u_{1},\ldots ,u_{n}\right) \right\} $

. The ordering between the 2

n

binary

n

-tuple probabilities,

$\Pr\left\{ \left( u_{1},\ldots,u_{n}\right) \right\} $

, can be illustrated by a directed graph which “scales” them by decreasing order, the so-called intrinsic order graph. In this context, this paper provides a simple algorithm for iteratively drawing the intrinsic order graph, for any complex stochastic Boolean system and for any number

n

of independent random Boolean variables. The presentation is self-contained.

We analyse different evolutionary spatial games, in which the pressure of the environment is taken into account, using binary cellular automata. The agents are unconditional players: at each time step a given cell cooperates (play C) or defects (play D) against all its neighbours. The pressure of the environment is implemented by requiring a minimum score

U

min

, representing indispensable resources (nutrients, energy, revenues, etc.) for an individual to prosper. Therefore a cell, instead of evolving just by adopting the state of its most successful neighbour, also takes into account if the ”winner” gets a score above or below

U

min

. In the latter case it has a probability of adopting the opposite state. Besides the paradigmatic and widely used Prisoner’s Dilemma (PD), two other games are analysed: the Hawk-Dove (H-D), popular in biology, and the Stag Hunt (SH) that recently came into favour in social sciences. The effect of the environmental stress is particularly dramatic in the case of the PD: it allows the evolution of cooperation for payoff matrices where defection was the rule for simple unconditional strategy players. Finally, we discuss a more sophisticated model version in which the ordinary evolutionary recipe of copying the most successful neighbour is supplemented with a ”win-stay, lose-shift” criterion. This model variant, for a restricted region of the parameter space, produces critical scaling laws.

We say that a Cellular Automata (CA) is coalescing when its execution on two distinct (random) initial configurations in the same asynchronous mode (the same cells are updated in each configuration at each time step) makes both configurations become identical after a reasonable time. We prove coalescence for two elementary rules and show that there exists infinitely many coalescing CA. We then conduct an experimental study on all elementary CA and show that some rules exhibit a phase transition, which belongs to the universality class of directed percolation.

Elliptic Curve Cryptosystems (ECC) are in the spotlight due to their significantly smaller parameters. The most costly arithmetic operation in ECC is division, which is performed by multiplying the inverse of a multiplicand. On the other hand, Cellular Automata (CA) have attracted a lot of attention regarding their potential for various applications. Thus, this paper presents an EC-based hardware architectural model for division based on CA over Galois Field GF(2

n

). The proposed architectural model is highly regular, expandable, and it has reduced latency based on periodic boundary CA. The proposed architecture can be easily implemented into the hardware design of crypto-coprocessors.

In this work, it is shown that a wide class of nonlinear sequence generators, the so called interleaved sequence generators, can be modelled in terms of linear cellular automata. A simple modelling procedure based on the concatenation of automata has been derived. The cryptographic characteristics of the generated sequences (period, linear complexity, number of different sequences) have been also analyzed. The technique is very simple and can be applied to generators in a range of practical applications.

Preimages of cellular automata are observed. Their number is computed using simple matrix operations. Three algorithms for making a list of preimages are graphically presented using the de Bruijn diagram and its concatenated form: the preimage network.

Cellular automata display configurations that are constant in time. We implement a stochastic synchronization between the present configurations of the system and its precedent ones in order to search for these constant patterns. For most of the known evolution rules with complex behavior a dynamic competition among all the possible constant patterns is established and no stationary regime is reached. For the particular rule coded by the decimal number 18, a self-synchronization phenomenon can be obtained, even when strong modifications to the synchronization method are applied.

In the previous paper [1] we presented general methods for detecting the evolution and dynamics of any one of the 255 fuzzy cellular automata (FCA) and showed that the method was applicable to all but nine of the 255 FCA. The main result there was that the limiting behavior of these FCA is decidable, except possibly for these nine, for finite initial configurations in a homogeneous background of zeros. Only six of these nine so called

exceptional

CA namely, FCA 172, 184, 202, 216, 226, and 228, appear to be interesting enough to warrant separate study, the other three, namely FCA 204, 228, and 240 being trivial. In this paper we study the exceptional FCA 184, a cellular automaton that admits a continuum of fixed points, namely the interval [0,1]. This FCA is of interest because the general technique developed in [1] fails for the determination of its asymptotics. We show, in particular, that the asymptotic evolution of FCA 184 from any finite initial including random configuration of non-zero cells is decidable.

This paper describes a novel implementation of a two-dimensional Cellular Automaton (CA) by introducing a

dormant

state. An overview of the use of CA’s in the field of Artificial Life reveals that certain crucial aspects of biological realism have been sacrificed in favour of abstraction or have not been considered at all. Conway’s famous “Game of Life” model includes certain fundamental aspects of population dynamics, including the transition from living state to dead state. But even the simplest biological system consists of more stages than the binary state in the Game of Life. The aim of this research is to build an extended CA model of natural biological systems by introducing a dormant state and to investigate the effect of dormancy on simple population dynamics.

The Dynamic Data Driven Application Systems (DDDAS) concept entails the ability to incorporate dynamically data into an executing application simulation, and in reverse, the ability of applications to dynamically steer measurement processes. Such dynamic data inputs can be acquired in real-time on-line or they can be archival data. DDDAS offers the promise of improving modeling methods, augmenting the analysis and prediction capabilities of application simulations, improving the efficiency of simulations and the effectiveness of measurement systems.

The scope of the present workshop provides examples of research and technology advances in enabling the DDDAS capabilities.

Previous work in the Instrumented Oil-Field DDDAS project has enabled a new generation of data-driven, interactive and dynamically adaptive strategies for subsurface characterization and oil reservoir management. This work has led to the implementation of advanced multi-physics, multi-scale, and multi-block numerical models and an autonomic software stack for DDDAS applications. The stack implements a Grid-based adaptive execution engine, distributed data management services for real-time data access, exploration, and coupling, and self-managing middleware services for seamless discovery and composition of components, services, and data on the Grid. This paper investigates how these solutions can be leveraged and applied to address another DDDAS application of strategic importance – the data-driven management of Ruby Gulch Waste Repository.

We describe how we plan to convert a traditional data collection sensor and ocean model into a DDDAS enabled system for identifying contaminants and then reacting with different models, simulations, and sensing strategies in a symbiotic manner. The sensor is just as useful in water as it would be on Mars for material identification. A successful terrestrial application of the sensor will lead to many new applications of the device and possible technology transfer to the private sector.

Threat management in drinking water distribution systems involves real-time characterization of any contaminant source and plume, design of control strategies, and design of incremental data sampling schedules. This requires dynamic integration of time-varying measurements along with analytical modules that include simulation models, adaptive sampling procedures, and optimization methods. These modules are compute-intensive, requiring multi-level parallel processing via computer clusters. Since real-time responses are critical, the computational needs must also be adaptively matched with available resources. This requires a software system to facilitate this integration via a high-performance computing architecture such that the measurement system, the analytical modules and the computing resources can mutually adapt and steer each other. This paper describes the development of such an adaptive cyberinfrastructure system facilitated by a dynamic workflow design.

Next-generation wireless sensor networks may revolutionize understanding of environmental change by assimilating heterogeneous data, assessing the relative value and costs of data collection, and scheduling activities accordingly. Thus, they are dynamic, data-driven distributed systems that integrate sensing with modeling and prediction in an adaptive framework. Integration of a range of technologies will allow estimation of the value of future data in terms of its contribution to understanding and cost. This balance is especially important for environmental data, where sampling intervals will range from meters and seconds to landscapes and years. In this paper, we first describe a general framework for dynamic data-driven wireless network control that combines modeling of the sensor network and its embedding environment, both in and out of the network. We then describe a range of challenges that must be addressed, and an integrated suite of solutions for the design of dynamic sensor networks.

We describe a prototype emergency response system. This dynamic data driven application system (DDDAS) uses wireless call data, including call volume, who calls whom, call duration, services in use, and cell phone location information. Since all cell phones (that are powered on) maintain contact with one or more local cell towers, location data about each phone is updated periodically and available throughout the cellular phone network. This permits the cell phones of a city to serve as an ad hoc mobile sensor net, measuring the movement and calling patterns of the population. Social network theory and statistical analysis on normal call activity and call locations establish a baseline. A detection and alert system monitors streaming summary cell phone call data. Abnormal call patterns or population movements trigger a simulation and prediction system. Hypotheses about the anomaly are generated by a rule-based system, each initiating an agent-based simulation. Automated dynamic validation of the simulations against incoming streaming data is used to test each hypothesis. A validated simulation is used to predict the evolution of the anomaly and made available to an emergency response decision support system.

A project concerned with applying Dynamic Data Driven Application Simulations (DDDAS) to monitor and manage surface transportation systems is described. Building upon activities such as the Vehicle-Infrastructure Integration initiative, a hierarchical DDDAS architecture is presented that includes coupled in-vehicle, roadside, and traffic management center simulations. The overall architecture is described as well as current work to implement and evaluate the effectiveness of this approach for a portion the Atlanta metropolitan area in the context of a hypothesized emergency evacuation scenario.

A Dynamic Data Driven Application System (DDDAS) was created to study interaction between fire and agent models during a fire evacuation. The analysis from that research can be used to validate proposed ideas in evacuation and building designs to ensure safety of buildings given various agent behaviors. Two separate models were used to simulate the components of the emergency situation: fire and agent. The independent models were able to run using data computed by the other interacting models, allowing careful examination of real-time interactions in a situation. Through study of the interactions, a better understanding is gained of how individual variables such as exit position and width affect the evacuation process and escape rate in the given scenario. Exits can be relocated and changed to quickly assess the effect on the model. The results can be used for improving building design and regulations as well as training first responders.

The total replacement value of the US transmission lines alone (excluding land) is conservatively estimated at over $100 billion dollars [1] and triples when including transformers and circuit breakers. Investment in new transmission equipment has significantly declined over the past 15 years. Some of the equipment is well beyond intended life, yet is operated under increasing stress, as load growth, new generation, and economically motivated transmission flows push equipment beyond nameplate limits. Maintaining acceptable electric transmission system reliability and delivering electric energy at low energy prices requires innovations in sensing, diagnostics, communications, data management, processing, algorithms, risk assessment, decision-making (for operations, maintenance, and planning), and process coordination. This paper overviews a comprehensive approach to develop methods and processes in these areas, driven by the ultimate objective to develop a hardware-software prototype capable of auto-steering the information-decision cycles inherent to managing operations, maintenance, and planning of the high-voltage electric power transmission systems.

In operations, simulation and control of power systems, the presence of real-time data relating to system states can yield precise forecasts and can enable robust active control. In this research we are developing efficient and robust methods to produce “data enhanced” reduced order models and filters for large-scale power systems. The application that this paper focuses on is the creation of new data-driven tools for electric power system operation and control. The applications systems include traditional

SCADA

systems as well as emerging

PMU

data concentrators. A central challenge is to provide near real-time condition assessment for ”extreme events,” as well as long-term assessment of the deterioration of the electrical power grid. In order to provide effective guidance for power system control, we are also developing visualization methods for integrating multiple data sets. These visualization methods provide an up-to-date view of the system state, and guide operator-initiated power system control.

This paper outlines the initial motivations and implementation scope supporting a dynamic data driven application system for material and structural health monitoring as well as critical event prediction. The dynamic data driven paradigm is exploited to promote application advances, application measurement systems and methods, mathematical and statistical algorithms and finally systems software infrastructure relevant to this effort. These advances are intended to enable behavior monitoring and prediction as well as critical event avoidance on multiple time scales.

Structural sensing and control is an important application of the DDDAS paradigm. Our work on structural sensing and control has several key aspects, including model reduction, control, simulation, and validation. Motivated by our work on validation using an actual three-storeyed structure, we are developing a comprehensive systems environment, Omni, for macroprogramming sensor networks. While there have been efforts targeted at enabling programmers to write lean applications for individual sensor nodes, there have been few efforts targeted towards allowing programmers to program entire networks as distributed ensembles. Omni provides an intuitive and efficient programming interface, along with operating system services for mapping these abstractions into the underlying network. In this paper, we provide a high-level overview of the Omni architecture, its salient features, and implementation details. The Omni architecture is designed to be a flexible, extensible, scalable, and portable system, upon which a wide variety of DDDAS applications can be built.

A Dynamic Data Driven Application Systems (DDDAS) approach is developed for evaluation of fluid-thermal systems wherein a complete specification of the boundary conditions is not known

a priori

and experimental diagnostics are restricted to a limited region of the flowfield. The methodology is applied to the configuration of a heated jet injected into a laminar boundary layer where the jet temperature is not known

a priori

. Preliminary results are presented.

We are interested in a DDDAS problem of localization of airborne contaminant releases in regional atmospheric transport models from sparse observations. Given measurements of the contaminant over an observation window at a small number of points in space, and a velocity field as predicted for example by a mesoscopic weather model, we seek an estimate of the state of the contaminant at the begining of the observation interval that minimizes the least squares misfit between measured and predicted contaminant field, subject to the convection-diffusion equation for the contaminant. Once the “initial” conditions are estimated by solution of the inverse problem, we issue predictions of the evolution of the contaminant, the observation window is advanced in time, and the process repeated to issue a new prediction, in the style of 4D-Var. We design an appropriate numerical strategy that exploits the spectral structure of the inverse operator, and leads to efficient and accurate resolution of the inverse problem. Numerical experiments verify that high resolution inversion can be carried out rapidly for a well-resolved terrain model of the greater Los Angeles area.

We consider a data assimilation technique for coupled ionospheric and thermospheric dynamics. The Global Ionosphere-Thermo-sphere Model (GITM) is used to simulate the ionospheric and thermospheric dynamics, and evaluate the performance of the data assimilation scheme that estimates the ion densities and flow speeds. This estimation technique is based on the state dependent Riccati equation (SDRE), which uses a frozen linear dynamics matrix for the time update of the error covariance and the evaluation of the Kalman filter gain. We demonstrate the performance of the data assimilation technique on a section of the ionosphere.

Classical formulations of data-assimilation perform poorly when forecast locations of weather systems are displaced from their observations. They compensate position errors by adjusting amplitudes, which can produce unacceptably “distorted” states, adversely affecting analysis, verification and subsequent forecasts. It is non-trivial to identify sources of position error, but correcting misplaced forecasts is essential for operationally predicting strong, localized weather events such as tropical cyclones. In this paper, we propose a method that accounts for both position and amplitude errors. The proposed method assimilates observations in two steps. The first step is

field alignment

, where the current model state is aligned with observations by adjusting a continuous field of local displacements, subject to certain constraints. The second step is amplitude adjustment, where contemporary assimilation approaches are used. Our method shows improvements in analyses, with sparse and uncertain observations.

Recently Support Vector Machines (SVMs) have played a leading role in pattern classification. SVMs are quite effective to classify static data in numerous applications. However, the use of SVMs in dynamically data driven application systems (DDDAS) is somewhat limited. This motivates the development of incremental approaches to handle DDDAS. In an incremental learning approach, it is critical to keep a certain number of support vectors (SVs) without seriously sacrificing the generalization performance of SVMs. In this paper a novel incremental SVM method, called an incremental revised support vector machine with filters (IRSVMF) is proposed to resolve the above limitations. Computational experiments with tornado data show that this approach is quite effective to reduce the number of SVs and computing time and to increase the detection rate of tornados.

Observing systems facilitate scientific studies by instrumenting the real world and collecting corresponding measurements, with the aim of detecting and tracking phenomena of interest. A wide range of critical environmental monitoring objectives in resource management, environmental protection, and public health all require distributed observing systems. The goal of such systems is to help scientists verify or falsify hypotheses with useful samples taken by the stationary and mobile units, as well as to analyze data autonomously to discover interesting trends or alarming conditions. In our project, we focus on a class of observing systems which are

embedded

into the environment, consist of

stationary and mobile

sensors, and

react

to collected observations by reconfiguring the system and adapting which observations are collected next. In this paper, we give an overview of our project in the context of a marine biology application.

We report on an ongoing effort to build a Dynamic Data Driven Application System (DDDAS) for short-range forecast of weather and wildfire behavior from real-time weather data, images, and sensor streams. The system changes the forecast as new data is received. We encapsulate the model code and apply an ensemble Kalman filter in time-space with a highly parallel implementation. In this paper, we discuss how we will demonstrate that our system works using a DDDAS testbed approach and data collected from an earlier fire.

The goal of this project is to develop a dynamic data-driven planning and control system for laser treatment of cancer. The research includes (1) development of a general mathematical framework and a family of mathematical and computational models of bio-heat transfer, tissue damage, and tumor viability, (2) dynamic calibration, verification and validation processes based on laboratory and clinical data and simulated response, and (3) design of effective thermo-therapeutic protocols using model predictions. At the core of the proposed systems is the adaptive-feedback control of mathematical and computational models based on a posteriori estimates of errors in key quantities of interest, and modern Magnetic Resonance Temperature Imaging (MRTI), and diode laser devices to monitor treatment of tumors in laboratory animals. This approach enables an automated systematic model selection process based on acceptance criteria determined a priori. The methodologies to be implemented involve uncertainty quantification methods designed to provide an innovative, data-driven, patient-specific approach to effective cancer treatment.

This paper presents a case of interdisciplinary collaboration in building and using a set of tools to compute the flows in a branched artery, to compare them with prior physical flow visualization, and to interpret them with further users in mind. The geometry was taken for a typical epicardial coronary artery with a side branch. The incompressible Navier-Stokes equations were solved with the hybrid spectral/hp element solver Nektar. Some simulations were visualized in the CAVE, an immersive 3D stereo display environment, and selected features are described and interpreted.

Great potential exists for future Brain Machine Interfaces (BMIs) to help paralyzed patients, and others with motor disabilities, regain (artificial) motor control and autonomy. This paper describes a novel approach towards the development of new design architectures and research test-beds for advanced BMIs. It addresses a critical design challenge in deriving the functional mapping between the subject’s movement intent and actuated behavior. Currently, adaptive signal processing techniques are used to correlate neuronal modulation with known movements generated by the subject. However, with patients who are paralyzed, access to the individual’s movement is unavailable. Inspired by motor control research, this paper considers a predictive framework for BMI using multiple adaptive models trained with supervised or reinforcement learning in a closed-loop architecture that requires real-time feedback. Here, movement trajectories can be inferred and incrementally updated using instantaneous knowledge of the movement target and the individual’s current neuronal activation. In this framework, BMIs require a computing infrastructure capable of selectively executing multiple models on the basis of signals received by and/or provided to the brain in real time. Middleware currently under investigation to provide this data-driven dynamic capability is discussed.

We present a dynamic data-driven framework for tracking gestures and facial expressions from monocular sequences. Our system uses two cameras, one for the face and one for the body view for processing in different scales. Specifically, and for the gesture tracking module, we track the hands and the head, obtaining as output the blobs (ellipses) of the ROIs, and we detect the shoulder positions with straight lines. For the facial expressions, we first extract the 2

D

facial features, using a fusion between KLT tracker and a modified Active Shape Model, and then we obtain the 3D face mask with fitting a generic model to the extracted 2

D

features. The main advantages of our system are (i) the adaptivity, i.e., it is robust to external conditions, e.g., lighting, and independent from the examined individual, and (ii) its computational efficiency, providing us results off- and online with a rates higher than 20

fps

.

Artificial intelligence (AI) can contribute to the management of a data driven simulation system, in particular with regard to adaptive selection of data and refinement of the model on which the simulation is based. We consider two different classes of intelligent agent that can control a data driven simulation: (a) an autonomous agent using internal simulation to test and refine a model of its environment and (b) an assistant agent managing a data-driven simulation to help humans understand a complex system (assisted model-building). In the first case the agent is situated in its environment and can use its own sensors to explore the data sources. In the second case, the agent has much less independent access to data and may have limited capability to refine the model on which the simulation is based. This is particularly true if the data contains subjective statements about the human view of the world, such as in the social sciences.

For complex systems involving human actors, we propose an architecture in which assistant agents cooperate with autonomous agents to build a more complete and reliable picture of the observed system.

One of the intended consequences of utilizing simulations in dynamic, data-driven application systems is that the simulations will adjust to new data as it arrives. These adjustments will be difficult because of the unpredictable nature of the world and because simulations are so carefully tuned to model specific operating conditions. Accommodating new data may require adapting or replacing numerical methods, simulation parameters, or the analytical scientific models from which the simulation is derived. In this research, we emphasize the important role a scientist’s insight can play in facilitating the runtime adaptation of a simulation to accurately utilize new data. We present the tools that serve to capture and apply a scientist’s insight about opportunities for, and limitations of, simulation adaptation. Additionaly, we report on the two ongoing collaborations that serve to guide and evaluate our research.

Traditional approaches to software systems development such as using tools and modeling frameworks are appropriate for building individual object oriented or component based software. However they are not suitable for designing of flexible distributed enterprise systems and open environments. In recent years, service-oriented architecture (SOA) has been proposed as a suitable architecture for development of such systems. Most current approaches in employing SOA are tailored to specific domains and hence are not general purpose. Therefore, in order to gain the full benefits of such technology, a more effective general approach to modeling and designing these complex distributed systems is required. In this paper, we present a model-driven approach to SOA modeling and designing complex distributed systems. In this approach, first the PIM of the business system is derived and expressed in standard UML modeling constructs and then this PIM is transformed to the SOA-based PIM by some transforming tool. After the SOA-based PIM is obtained, it can be used to generate PSM for a specific platform such as Web Services, Jini or other platforms. To make it clear how this PSM could be generated we will use Web Services as a target platform and the steps of this transformation will be shown.

Most current 3D landscape visualisation systems either use bespoke hardware solutions, or offer a limited amount of interaction and detail when used in realtime mode. We are developing a modular, data driven 3D visualisation system that can be readily customised to specific requirements. By utilising the latest software engineering methods and bringing a dynamic data driven approach to geo-spatial data visualisation we will deliver an unparalleled level of customisation in near-photo realistic, realtime 3D landscape visualisation. In this paper we show the system framework and describe how this employs data driven techniques. In particular we discuss how data driven approaches are applied to the spatiotemporal management aspect of the application framework, and describe the advantages these convey.

For multi-channel process, due to disadvantages of the open-loop single variable step method, multi-channel test method is used. That means all of the channels are tested at the same time. In order to eliminate cross-effect of the different test signals, it requires that all the test signals are uncorrelated. Several test signals are introduced and analyzed. Based on two familiar identification methods: correlation analysis method and least-squares method, we put our strength on the way to get uncorrelated test signals. A novel design for the period length of uncorrelated pseudo random binary sequence (PRBS) is proposed. Use this design method, identifiable PRBS signals can be gained and their periods are the shortest. Simulation results show the effectiveness.

Efficient knowledge storage is for easy to look up and speed up the reaction. Knowledge resource library stored large amount of knowledge achieve, and it can be distill through all kinds of links. Technically, there are two concerns, one is to index through what kind of related strategy, establishes keyword controls, in order to apply the standard literature management tools, the other is how to adopt efficient storage strategies, in order to index and update the knowledge system fast, this thesis takes the operation process as a starting point, investigates the organization strategies towards knowledge classification process and storage, as well as provide a practice case based on research and develop of the steel product.

The software package, Small WebComputing (SWC), has been applied to a Monte Carlo simulation of a system of hard hyperspheres in a variety of dimensions. The SWC environment was chosen because once the framework is embedded in the application code, the user has the choice of running the distributed computations as a set of applets, as parallel threads on a symmetric multiprocessor or as independent processes distributed over a network. A description of the software and a discussion of its ongoing evolution is presented.

In this paper we present a Grid application developed for electron transport problems called

SALUTE

(

S

tochastic

AL

gorithms for

U

ltra-fast

T

ransport in s

E

miconductors). We consider a physical model of a femtosecond relaxation of optically excited electrons which interact with phonons in an one-band semicondoctor. The electron-phonon interaction is switched on after a laser pulse creates an initial electron distribution. The Barker-Ferry equation is utilized as a quantum-kinetic model of the process under consideration. Two cases of this process are investigated – with and without an applied electric field. The electric field causes shift in the replicas, population of the semiclassically forbidden regions and influences the broadening and retardation of the electron distribution. The paper describes Grid implementation of these CPU-intensive algorithms. Using this application innovative results for different materials can be obtained. Here we present the first version of SALUTE which is used to obtain innovative results for GaAs materials. The results from a number of tests on MPI-enabled Grid are shown and disscussed.

This paper considers the joint state and parameter estimation of extended targets. Both the target kinematic states, position and speed, are estimated with the target extent parameters. The developed algorithm is applied to a ship, whose shape is modelled by an ellipse. A Bayesian sampling algorithm with finite mixtures is proposed for the evaluation of the extent parameters whereas a suboptimal Bayesian interacting multiple model (IMM) filter estimates the kinematic parameters of the maneuvering ship. The algorithm performance is evaluated by Monte Carlo comparison with a particle filtering approach.

In this paper we present error analysis for a Monte Carlo algorithm for evaluating bilinear forms of matrix powers. An almost Optimal Monte Carlo (MAO) algorithm for solving this problem is formulated. Results for the structure of the probability error are presented and the construction of robust and interpolation Monte Carlo algorithms are discussed.

Results are presented comparing the performance of the Monte Carlo algorithm with that of a corresponding deterministic algorithm. The two algorithms are tested on a well balanced matrix and then the effects of perturbing this matrix, by small and large amounts, is studied.

In this paper we consider bilinear forms of matrix polynomials and show that these polynomials can be used to construct solutions for the problems of solving systems of linear algebraic equations, matrix inversion and finding extremal eigenvalues. An almost Optimal Monte Carlo (MAO) algorithm for computing bilinear forms of matrix polynomials is presented.

Results for the computational costs of a balanced algorithm for computing the bilinear form of a matrix power is presented, i.e., an algorithm for which probability and systematic errors are of the same order, and this is compared with the computational cost for a corresponding deterministic method.

Recently, a main memory index structure called the cache conscious-generalized search tree (CC-GiST) was proposed. The CC-GiST is such a novel index structure that it can be used for implementing all the existing cache conscious trees with the minimal efforts. It incorporates the pointer compression and the key compression techniques, which were adopted by the existing cache conscious trees to reduce the cache misses, in a single framework. In this paper, we formally analyze the performance of the CC-GiST. We compare the performance of the CC-GiST with the existing cache conscious trees. The result shows that the CC-GiST has the negligible overhead for supporting all the existing cache conscious trees in a single framework, and the performance of the tree is almost unaffected.

Redo log of database must be written to permanence storage like disks. When database is heavily loaded, the crowded redo log writing queue will become a performance bottleneck. In this paper, the operation principle of database redo log is analyzed. It is found that if redo log is stored on virtual memory disk directly, the database will demonstrate better performance. In addition, the reliability of redo log system based on virtual memory disk is analyzed. At the end of the paper, a contrastive performance measurement result is given.

In out-of-band virtualization systems, it is typical that the virtual storage manager (VSM) maintains the metadata for all Agents and performs the address mapping. The Agent performs I/O access according to the mapped address. High network latency, poor performance and low scalability are the main problems of this method. This article introduces an improved design and implementation of an out-of-band virtualization system, in which each Agent maintains a copy of the metadata and performs the address mapping by itself, so as to improve performance, reliability and scalability. This article discusses the on-line extension of a logical volume and the related problems such as synchronization. This function improves the ability of uninterrupted service. The system was tested on the UFS of Solaris 10 and the results showed that the performance of a stripe volume consisting of four disks exceeded that of a linear volume by an average of 103.65%.

Built on a sequential write/read device, a tape library is seldom considered as a viable place for fast backup/restore data. With the help of the virtualization technology, in this paper we propose a virtual backup and archive system, called VBAS. The purpose of VBAS is to maintain a consistent view of mass storage so that the user can effectively manage it. And VBAS allows users to create files and directories as well as delete, open, close, read, write and/or extend the files on the device(s). VBAS maintains security on the files and provides the management for fragmentation. Moreover, VBAS can support large-scale file systems. Users have two ways to access VBAS: using general backup application, and through the APIs provided by VBAS. Based on VTL, RAID-DP, and iSCSI, VBAS not only has the disk-file-system-like functions, but also retains the characteristics of tape library storage, thus achieving a good tradeoff between cost and performance. The prototype system performance is presented and improvements are analyzed to achieve higher write/read performance.

Digital storage is a key element not only of computing systems, but is now considered an essential component of the infrastructure of any modern organization. This need has co-evolved with the technology that has grown rapidly in recent years to provide low-cost high-capacity storage. At the same time, the storage needs of users have now become more sophisticated and diverse. Some users require very long-term preservation; others need high security; and still others ask for highly-reliable, distributed storage solutions. These needs pose a problem for solution providers in that no single solution seems to meet all needs. Similarly, users must construct services out of disk systems on their own. This paper proposes a way to streamline the marketplace through

insurable storage services

, a combination of two ideas. The first is to define different categories of storage service; the assumption here is that a refined categorization will better identify particular user needs. The second, and more substantive idea, is to treat digital documents as insurable property. The insurance of storage will provide economic incentives for both producers (storage service providers) and consumers (individuals, organizations) to jointly create a marketplace that provides a diversity of differentially-priced services. For example, insurers can help assess the durability of storage solutions and provide consumers with a quantitative valuation (“It’ll cost you $x per GB to ensure that your documents last 100 years”). Similarly, storage service providers will have incentives to maintain multiple geographically distributed copies, and to continually move the copies onto emerging technologies (“You’ll need to store more copies if you want a higher reliability rating”).

The Object-based storage is an emerging storage architecture that could easily fulfill multi-dimensional storage QoS requests. This paper focuses on providing QoS guarantees under Object storage infrastructure along the three most prevalent dimensions: capacity, bandwidth and latency through storage resource allocation and IO commands scheduling. Firstly we propose an algorithm on storage resource mapping derived from Toyoda algorithm, which achieves efficient resource utilization through consideration of the OSDs’ serving ability. Secondly we propose an object commands scheduling mechanism and develop a prototype system based on the Lustre filesystem. Through adding timestamp to each object command and scheduling the command queue by final finish time, the system can efficiently fulfill the demands on latency from the front applications.

As storage system scales to thousands of disks, data distribution, load balance and the support for heterogeneous disks become increasingly important. In this paper, we present a new data-placement method named Weighted Interval Algorithm (WIA) for heterogeneous disks. Through it is not optimal in some circumstances, the difference between WIA and the optimal algorithm is trivial. Combined with replication, WIA can nearly balance access load and space utilization and improve reliability simultaneously. For the first time, we implement a data-placement system with high scalability, reliability and performance. The experimental results show that WIA reduces the average response time by 14.8% and decreases coefficient of relative load from 78.09% to 47.46% while the difference of the ratio of space utilization between disks is not more than 0.79%.

In this paper symbolic, supervised learning is used in a multi-agent system for resource management. Environment is a Fish Bank game, where agents manage fishing companies. Rule induction is applied to generate ship allocation and cooperation rules. In this article system architecture and learning process are described and experimental results comparing performance of several types of agents are presented. The results obtained confirm that applying a supervised learning algorithm in a multi-agent system may improve resource management.

The main purpose of this article is to present multi-criteria decision-making principle as a tool providing the agent with autonomous decision-making capacity. The advantages and disadvantages of this principle are described on the example of the robot soccer game, together with the future perspective of this approach. The control system for robot soccer game is designed with consideration of using a multi-criteria decision-making. As a quickly changing environment, the robot soccer game provides an excellent testing ground for such experimental approach.

The paper proposes a JADE-based A-Team environment (JADE-A-Team) as a middleware supporting the construction of the dedicated A-Team architectures used for solving variety of computationally hard optimization problems. The paper includes a general overview of the functionality and structure of the proposed environment and a more detailed description of optimization agents including their standard functions, ontology, construction requirements and activation procedure. Further sections explain how to create and activate an A-Team agent and how the communication between agents is handled. Conclusions focus on advantages of the JADE-A-Team environment and on suggestions for further research.

Ambient Intelligence represents a vision of the future whereby the world will be saturated with embedded electronic devices that are sensitive and responsive to people. This technology will combine the concepts of intelligent systems with that of pervasive computing. Intelligent agents of varying capabilities will provide the foundations for many applications within this domain. As a means of achieving this objective a framework – Agent Factory Micro Edition (AFME) has been developed to enable the creation of agent-based applications on computationally constrained devices such as cellular digital mobile phones. It has been specifically designed to tackle the performance and memory footprint issues associated with executing intentional agents on mobile devices.

High degree of complexity of processes connected with compound chains of operations requires search for new models and methods. Moreover, one should take into account possibilities of presence of undesired situations, called crises, when system functions in real conditions. The aim of the work is to create a model of a system based on application of agents, which manages a process of realization of chains of operations (either a "matter" to be processed or production process) and its structure in such a way that it is possible to syn chronise the model by real process. In the paper, a concept of multiagent system solving this problem, concerning potential crisis situations, as well as prototype realisation for a sample application have been shown.

The paper presents an analysis of requirements for building simulation systems with tightly coupled components, such as typical mesh based PDE approximation software. The considered systems are characterized as having high communication to computation ratio. When designing architectures for such systems the hardware and middleware capabilities for providing communication links between processes have to be investigated and fully exploited. This is the place where agent technology perfectly fits the requirements. In the whole system, the capabilities of agents should be complemented with less flexible but more efficient software organization.

As an example a framework for finite element simulations, employing a modular architecture (described in [1]), is considered. Communication requirements for typical computations are estimated and evaluated in view of possible inter-process communication. The role of agents in setting up the execution structure of simulations is described.

In the

hp

adaptive Finite Element Method (FEM) applications, the computational mesh consists in finite elements with varying size

h

, and varying polynomial order of approximation

p

on finite element edges, faces and interiors. The parallel

hp

adaptive codes work on the computational domain partitioned into sub-domains with each of the sub-domains delegated to a single processor. The algorithm of parallel mesh refinements on such a distributed FE must enforce global mesh regularity rules. The paper presents the applications of multiple agents to implement the parallel mesh refinements algorithm. Agents work on distributed data structure storing FE mesh where dynamic mesh refinements are recorded by growing trees of initial mesh elements nodes. Agents located on separated sub-domains communicate in order to establish necessary actions on the distributed mesh.

Along with the evolution of the numerical methods new software methodologies have been developed. Goal of our research is to apply Aspect Oriented Programming (AOP) in the development of multiagent simulation system. In this paper theoretical model of aspect-multiagent system is presented, its architecture and implementation is described. Results of experiments performed conclude the work. The model considered here can serve as a design tool for foundry processes, especially to design conditions for cooling of a casting leading to desired crystal structure.

This work concerns the design and development of a driving simulation system, which exhibits intelligent driving behaviour at the tactical level, as part of a traffic simulation environment. Our tactical driving system using genetic algorithms, named GA-INTACT, accounts for the subject vehicle and other vehicles positions and speed parameters in the surrounding traffic condition, and selects favourable speed change and lane transition actions for the ‘subject’ vehicle, according to safety, speed and driving behaviour criteria. Simulation results demonstrated that the adoption of the Genetic Algorithms approach for obtaining near-optimum driving solutions eliminates the need for learning driving patterns, and allows the efficient handling of the complex nature of tactical driving modelling problem. The role of the driving behaviour in influencing the outcome of the driver’s decision is emphasised, an aspect that was not treated sufficiently in previous tactical driving simulation approaches.

The problem of mobile robot navigation has received a noticeable attention over last few years. Several different approaches were presented, each having major limitations. In this paper a new, agent-based solution the problem of mobile robots navigation is proposed. It is based on a novel representation of the environment, that divides it into a number of distinct regions, and assigns autonomous software Space Agents to supervise it. Space Agents create a graph, that represents a high-level structure of the entire environment. The graph is used as a virtual space, that robot controlling agents work in. The most important features of the approach are: path planning for multiple robots based on most recent data available in the system, automated collision avoidance, simple localization of a ”lost robot” and unrestricted scalability.

The paper presents the Agent-Oriented technology for running the parallel CAE computation. Fast and effective distributed diffusion scheduling is available that minimizes computation and communication time necessary for task governing and provides transparency in resource availability. Detailed evaluation of the diffusion rule parameters was obtained in the course of analysis of computational, memory and communicational complexity of CAE tasks.

We propose a dynamic resource allocation mechanism which can be used in multi-agent computer network interconnection management systems. Considering a setting of multiple consumers and elastic supply we argue that interaction between autonomous system resource managers is a game. We make the following contributions. First, we analyze the stability of the Nash equilibrium point of the resource allocation game. Second, we show that in comparison to the Cournot mechanism the mechanism we propose may lead to solutions which are characterized by a larger aggregate surplus.

The paper extends the formal model of a computing Multi Agent System introduced in our previous papers to the case in which the background load coming from operating systems activities as well as other applications is included. Results concerning the existence of the optimal scheduling strategy as well as the characterization of such strategies have been obtained. The theorems partially verify scheduling heuristics (diffusion rules) designed and tested for large scale CAE computations.

This paper presents a fault-tolerance scheme based on mobile agents for the reliable mobile computing systems. The mobility of the agent is suitable to trace the mobile hosts in the system and the intelligence of the agent makes it efficient to support the fault-tolerance services. The proposed scheme especially focuses on the adaptiveness of the agent. The agents try to keep the failure-recovery cost and the failure-free operation cost within a certain level, regardless of the hand-off frequency of the mobile hosts. They also try to balance the two costs.

In this paper, we evaluate the feasibility of multiagent control of resources to be shared in user networks. A user network is totally controlled by the users, both at application and transport level. One of the possible applications in these networks is peer-to-peer (P2P) file exchange sharing the "external" access to the Internet (set of links between the user network and the Internet). If a node cannot serve its demand with its own external link, it requests help from another node via the high-bandwidth internal user network. We model user nodes as agents to simulate and to evaluate a new agent-based distributed control scheme. The simulation results in this paper confirm that it is possible to improve resource sharing in user networks using agents that take decisions autonomously, from local information, and check that file exchange services offered to neighbour nodes do not surpass appropriate credit limits.

Ethically–social approach to security problem uses the idea of agents functioning evaluation in a multi–agent system in the analogous way to the evaluation of a person’s behavior in small human societies. This approach involves distributed evaluations made by autonomous agents and processing of the results of this evaluations in order to create a collective decision of a society of agents. Research presented in this paper focuses on the part of domain of ethically–social behavior evaluation that specifies how an agent evaluates the behavior of another agent. The idea of heterogeneous behavior evaluations is formulated and tested in this domain.

The paper presents some modification of the idea of an evolutionary multi-agent system for multiobjective optimization, dealing simultaneously with the stagnation of evolutionary process and the loss of agents representing high-quality solutions. The main mechanisms proposed follow the idea of

elitist operators

known from classical evolutionary algorithms, yet in this case the

elite

does not take part in the evolutionary process. Some preliminary results based on a typical multi-objective problem presenting the most important features of the proposed approach are also discussed.

The basic idea of the approach proposed in this paper is to apply multi-agent paradigm in order to enable the integration and co-operation of different knowledge acquisition and representation techniques. The effective operation of learning process is achieved by evolutionary optimization running at the level of agents’ population. In the discussed variant of the model, each agent uses reinforcement learning, and the obtained knowledge is represented as the set of simple decision rules. The approach is illustrated by a particular realization of the system dedicated to the evasive maneuvers problem, together with preliminary experimental results.

Premature convergence, the major problem that confronts evolutionary algorithms, is also encountered with the Particle Swarm Optimization (PSO) algorithm. In the previous work [11], [12], [13], the Quantum-behaved Particle Swarm (QPSO) is proposed. This novel algorithm is a global-convergence-guaranteed and has a better search ability than the original PSO. But like other evolutionary optimization technique, premature in the QPSO is also inevitable. In this paper, we propose a method of controlling the diversity to enable particles to escape the sub-optima more easily. Before describing the new method, we first introduce the origin and development of the PSO and QPSO. The Diversity-Controlled QPSO, along with the PSO and QPSO is tested on several benchmark functions for performance comparison. The experiment results testify that the DCQPSO outperforms the PSO and QPSO.

Application of mechanisms of immune memory in the computer security domain allows to increase performance of certain class of security systems that are based on detection of attacks without

a priori

knowledge of attack’s technique. Immune memory should enable the system to memorise once encountered attacks and prevent it together with its consequences in the future. The use of agent technologies gives new possibilities in the management of stored attack’s patterns — patterns of obsolete attacks should be deleted but those of new and frequent should be maintained and generalised. In this paper ideas from agent technology and immune memory domain are introduced into computer security, tested and discussed.

In this paper, an improved negative selection procedure to handle constraints in a multi-criterion evolutionary algorithm has been proposed. The problem that is of interest to us is the complex task assignment for a distributed computer system. Both a workload of a bottleneck computer and the cost of system are minimized; in contrast, a reliability of the system is maximized. Moreover, constraints related to memory limits and computer locations are imposed. Finally, an evolutionary algorithm with tabu search procedure and the improved negative selection is proposed to provide effective solutions.

Co-evolutionary techniques for evolutionary algorithms are aimed at overcoming their limited adaptive capabilities and allow for the application of such algorithms to problems for which it is difficult or even impossible to formulate explicit fitness function. In this paper the idea of

co-evolutionary multi-agent system with host-parasite mechanism for multi-objective optimization

is introduced. In presented system the Pareto frontier is located by the population of agents as a result of co-evolutionary interactions between species. Also, results from runs of presented system against test functions are presented.

The mathematical model of multi agent resource conversion processes (RCP) is developed by the means of discrete-event simulation systems and expert systems. Within the framework of mathematical model RCP are defined: production system of the RCP structure, that taking into account conflicts origin. The discrete-event simulation system "BPsim" is developed on the basis of the multi agent RCP mathematical model. The “BPsim” system is inculcated on the firms in Ural region (Russia).

The paper deals with visual computational design in which emergence is a key to creativity. The presented framework for conceptual design uses shape grammars and curious agent assistants. The intelligent agents perceive the changing environment and emergent phenomena that occur in it. Interacting with each other and the designer they look for the most original and plausible solutions to a given design task. The approach is illustrated by the example of a designing floor-layouts.

Agent systems demand from graph transformation systems not only an effectiveness of parsing but also supporting of the online derivation control (in order to visualize or control of their development). The way of the effective realization Derivation Control Environment for the distributed systems is presented. Finally, the computational effectiveness of the parsing, membership problems of the full graph transformation system is discussed.

In this paper a strategy for integration of agents’ prepositional attitudes is proposed in order to solve the semantic inconsistency between opinions of members of agents’ population. In proposed algorithm all the agents’ answers are transformed into fuzzy set equivalents and then final formula representing the agreement of all opinions is obtained.

Service discovery is becoming a hot topic as it becomes more widespread through the Internet. In pervasive environments, there are various services, and users use a service discovery protocol for finding the most appropriate service among them. Thus, over the past few years, many service discovery protocols have been proposed. Even though they provide convenient service discovery functionality with users, complexity is increased because they have different message exchange pattern and architecture. In this paper, we propose a novel middleware based on agent platform for interoperability among heterogeneous services. This has the Discovery Agent for each service discovery protocols and it can detect service advertisement messages for registration, so the services are registered in agent platform. Finally, user can use them. The key point of this is not to require modification of existing service discovery protocols. Near the conclusion of this paper, this is implemented.

There is a high interest on service discovery techniques, which minimize the cost of detecting services and provide users with convenience. For more dynamic and useful service discovery, middleware for context-aware service discovery is required. This paper proposes an intelligent middleware architecture for context-aware service discovery that is based on JADE, which is a FIPA-compliant agent platform. The proposed middleware provides not only service scalability among heterogeneous domains, but also modules processing context information. When a service is detected, context information relating to the user and environment is used. As a policy-based system, the proposed middleware also use predefined policy. In other words, user preference can be considered. Near the conclusion of this paper, we discuss future works.

This paper introduces a distributed publication alerting service which increases the amount of information in notification message while using information hiding principle. It is aimed to design a mobile agent based publication alerting system (MABPAS) which uses mobile agents to dispatch notification information (generally all text information) about produced publication to subscribers. MABPAS combines the advantages of publish/subscribe communication model and mobile agents into a flexible and extensible distributed execution environment.

Niching techniques for evolutionary algorithms are aimed at maintaining the diversity through forming subpopulations (species) in multi-modal domains. Similar techniques may be applied to evolutionary multi-agent systems, which provide a decentralised model of evolution. In this paper a specific EMAS realisation is presented, in which the new species formation occurs as a result of co-evolutionary interactions between preexisting species. Experimental results aim at comparing the approach with a classical niching techniques and a basic EMAS implementation.

Many geospatial models are developed using command line modules of GIS packages. To utilize scientific workflow technology in geospatial modeling, it is important to support command line GIS modules in scientific workflow systems. However, straightforward representation of command line modules as workflow components conflicts with conventional conceptual design patterns. We propose a two-step geospatial scientific workflow composition approach. Simple conceptual workflows are composed in the first step. These allow data type-based workflow validation. The validated conceptual workflows are then transformed automatically into executable workflows using command line modules in the second step. We describe the preliminary implementation of the proposed approach in the Kepler scientific workflow system and demonstrate its feasibility using an example.

Emerging Grid technologies enable solving scientific problems that involve large datasets and complex analyses. Coordinating distributed Grid resources and computational processes requires adaptable interfaces and tools that provide a modularized and configurable environment for accessing Grid clusters and executing high performance computational tasks. In addition, it is beneficial to make these tools available to the community in a unified framework through a shared

cyberinfrastructure

, or a portal, so scientists can focus on their scientific work and not be concerned with the implementation of the underlying infrastructure. In this paper we describe a scientific workflow approach to coordinate various resources as data analysis pipelines. We present a three tier architecture for LiDAR interpolation and analysis, a high performance processing of point intensive datasets, utilizing a portal, a scientific workflow engine and Grid technologies. Our proposed solution is available through the GEON portal and, though focused on LiDAR processing, is applicable to other domains as well.

Aerodynamicists use wind tunnels to aid the research, design and development of products such as aircraft, cars, and yachts, amongst others. The data acquired from such tests must be acquired, collated, processed and analysed in a timely fashion to maximise productivity. In many scenarios a variety of data acquisition systems are used, and managing the overall testing process can be challenging. The wind tunnel grid project aims to provide an extensible, network-based system that can provide a more seamless working environment for scientists and engineers, so that they can focus on the data analysis and interpretation part of the process.

In this paper we describe the development and implementation of the wind tunnel grid system workflow. By exploiting Windows Workflow Foundation we are able to provide an easy-to-use and extensible workflow environment (wind tunnel grid workflow framework) that meets the requirements of both the developer and end-user well. By leveraging the .NET-based CoG Toolkit previously developed, interoperability with Globus grid services is demonstrated.

In this work we present a tool for manually connecting services by means of their input and output data, helping bioinformaticians in the automatic evaluation of workflows for the dynamic integration of heterogeneous data sources, services and computational resources. Our workflow platform offers a view of the different tools available as a single and uniform pool of services readily available for enhancing query processing.

Virtual Laboratory (VLab) has been developed as a distributed component system supporting a remote access to physical devices via the Internet. Unlike in typical VLabs, much attention has been paid to designing a state-of-the-art architecture which facilitates among others exposing the functionality of the devices, composing them into complex experiment stands as well as supervising in an interactive or automated way. The paper describes an application of workflow to experiment control in the VLab. Nevertheless, the presented concept consisting in wrapping a workflow engine by a component of a selected platform component has much more general nature.

The paper presents design, implementation details and simulations of scientific workflows involving compute-intensive tasks on clusters and PCs. The author has incorporated support for scientific workflows into previously developed J2EE-based BeesyCluster, deployed at Academic Computer Center Gdansk Poland on large HPC resources including a large 288-processor Itanium2 cluster. BeesyCluster allows users to manage various accounts on clusters/PCs via WWW/Web Services, run shell interactively, compile, queue, run tasks, publish services for other users, work in teams. A frequent scenario in HPC computing is analyzed, in which a workflow is combined from tasks offered by different users. Steps of the workflow include data preparation and following simulations run in parallel on clusters, with and without queuing systems.

Grid workflow, as a basic service in the grid environment, is a useful tool to help researchers make use of various gird resources to solve scientific problems. It is possible lots of users invoke grid workflow services in a very narrow time interval. Therefore, a centralized grid workflow engine is apt to be a bottleneck. In the paper, a novel grid workflow engine is proposed. It is based on Jini platform and employs Jini services to implement functions of grid workflow engine. By adding or removing enactment services, the grid workflow engine can be reconfigured dynamically. The workflow manager of the engine can allocate requests to proper enactment services according to some load-balancing strategy. The structure and mechanisms of this grid workflow engine are discussed in the paper. A prototype system and testing results are also introduced.

Many Grid architectures have been developed in recent years. These range from the large community Grids such as LHG and EGEE to single site deployments such as Condor. However, these Grid architectures have tended to focus on the single or batch submission of executable jobs. Application scientists are now seeking to manage and use physical instrumentation on the Grid, integrating these with the computational tasks they already perform. This will require the functionality of current Grid systems to be extended to allow the submission of entire workflows. Thus allowing the scientists to perform increasingly larger parts of their experiments within the Grid environment. We propose here a set of high level services which may be used on-top of these existing Grid architectures such that the benefits of these architectures may be exploited along with the new functionality of workflows.

The Imperial College e-Science Networked Infrastructure (ICENI) has been developed by the London e-Science Centre for over four years. ICENI has prototyped many novel ideas for providing an end to end Grid middleware. This has included: service-oriented architecture, component programming model, retaining and using meta-data collected throughout the life-cycle of an application, and scheduling algorithms which are aware of workflow and performance data. In this paper we describe the workflow pipeline and deployment process of ICENI II. This allows the user to specify their workflow at an abstract level which is fed through the pipeline in order to successfully deploy it over the Grid. The tool sets, which make up the stages of the ICENI II pipeline, are designed to be composable in an à-la-carte fashion. Thus allowing Grid developers to select only those components which are relevant for their work. When these tool sets are composed together they form the higher level services required to make the Grid useful to the end user.

In this paper, we propose an agent-based middleware architecture for the workflows used in Grid portal. We combined workflow model, workflow management system, job distribution, parameter scheduling, and service inteface in the middleware architecture. The workflow model consists of three layers. The agent-based middleware architecture consists of five agents and three types of communication protocols. Users can design a lightweight, flexible and effective middleware, which can be used to construct a workflow-based Grid portal.

The work described in this paper is a contribution to the problems of managing in data-intensive scientific applications. First, we discuss scientific workflows and motivate there use in scientific applications. Then, we introduce the concept of cooperative processes and describe their interactions and uses in a flexible cooperative workflow system called

Bonita

. Finally, we propose an approach to integrate and synthesize the data exchanged by the mapping of data-intensive science into Bonita, using a binary approach, and illustrate the endeavors done to enhance the performance computations within a dynamic environment.

In this paper we present design and development of a knowledge framework for grid and web service-based workflow composition and execution. We highlight the corresponding architecture and the process of service annotation, discovery and composition in the project K-WfGrid [5]. We describe in detail the challenges of a flood-forecasting application and corresponding design and development of the service oriented model, which is based on the well known Web Service Resource Framework (WSRF). Semantic descriptions of the WSRF services are presented as well as the architecture, which exploits semantics in discovery and composition of service workflows. Further, we demonstrate how experience management solutions can aid the process of collaborative service discovery and composition. The whole K-Wf Grid system provides a unique approach in Semantic Grids by combining the advances of semantic web services and grid architectures.

Quality (QOS) prediction is one of the most important research topics of workflow management system. In this paper, we propose the SWQ approach to analytically evaluate QOS of workflow systems based on QWF-net, which extends traditional WF-net by associating tasks with exponential response time and exponential TTF (time-to-failure). The comparison between simulative and analytical results in the case study indicates that the SWQ approach achieves satisfactory accuracy. The paper concludes with a comparison between the SWQ approach and other related work.

The service-oriented approach to performing distributed scientific research is potentially very powerful but is not yet widely used in many scientific fields. This is partly due to the technical difficulties involved in creating services and composing them into workflows. We present the Styx Grid Service, a simple system that wraps command-line programs and allows them to be run over the Internet exactly as if they were local programs. Styx Grid Services are very easy to create and use and can be composed into powerful workflows with simple shell scripts or more sophisticated graphical tools. Data can be streamed directly from service to service and progress can be monitored asynchronously using a mechanism that places very few demands on firewalls. Styx Grid Services can interoperate with Web Services and WS-Resources.

Different planning techniques have been proposed so far which address the problem of automated composition of web services. However, in realistic cases, the planning problem is far from trivial. Such issue is more serious when services are embraced in grid environments, which are usually resource-conscious. We propose a planning techniques for the automated composition of grid services described in OWL-S process models. The technique allows for the synthesis of plans that encode compositions of grid services with the usual programming constructs. We apply this technique in our DDGrid project and do some preliminary experimental evaluations.

This paper advocates a non-intrusive and incremental approach to enabling existing Grid programming systems with new features. In particular, it presents a software component enabling NetSolve applications with direct communications between remote tasks. The software component is a supplementary one working on the top of the basic NetSolve system. Its design also allows remote tasks to be freely mixed in a single application, independent on whether each particular task is enabled for direct communications or not. Experiments with this software are also presented.

The dynamic nature and the geographic distribution of scientific resources, require flexible and adaptive computational environment where an in-silico experiment can be executed as a workflow of activities. In this paper, we propose a software environment to dynamically generate domain-specific, agent-based workflow engines from workflow specifications. The workflow engine is a proactive multiagent system -a distributed, concurrent system- whose autonomous components interact in performing the workflow activities in a specific domain. The proposed approach has been implemented on Hermes, agent-based mobile computing middleware, and tested within “Oncology over Internet” project.

Properties of traffic noise and flow are often measured on complex networks and are used to diagnose the network’s functional state and underlying structure, even though the precise structure–function interdependences are often unknown. Here we attempt to unravel some basic interdependences between structure and traffic on networks in numerically controlled traffic models. We simulate constant-density traffic on two different network topologies, which emerge from the same preferential rewiring rules but one within growth and the other under static conditions. We determine universal noise properties and the maximal-flow spanning trees on these classes of network topologies. We study both low-density traffic (structure dependences) and high-density traffic, where queuing influences transport properties.

We introduce four algorithms for packet transport in complex networks. These algorithms use deterministic rules which depend, in different ways, on the degree of the node, the number of packets posted down each edge, the mean delivery time of packets sent down each edge to each destination and the time since an edge last transmitted a packet. On scale-free networks all our algorithms are considerably more efficient and can handle a larger load than the random walk algorithm. We consider in detail various attributes of our algorithms, for instance we show that an algorithm that bases its decisions on the mean delivery time jams unless it incorporates information about the degree of the destination node.

Genetic algorithm (GA) has been widely used in optimizing and solving various problems since first proposed, and its characters also have been deeply studied. In this paper, we investigate the benefits of genetic algorithm whose population is distributed on small-world networks. In particular, we pay our attention to the complexity of how small-world affects the behavior of spatial GA. Our work shows that, on a complex problem, the behavior of spatial GA on the small-world networks is influenced by at least two different factors: local selection and asymmetric topology. It is more complex than previous results from simple lattice models. Our results could provide lots of potential methods to improve the performance of spatial GA and give some guidance for designing of parallel genetic algorithm. We also present many future problems on the influence of small-world to spatial GA.

Instead of actual experiments to network protocols, network simulators are useful to analyze these network protocols for lower analysis cost. The Internet topology is dynamically evolving and growing, and then shows changing characteristics based on time flow. Studies of Internet topology have been motivated by the demands for analysis and simulation to the modeling of real networks. Hence, to develop the Internet simulator, proper characteristics to Internet topology should be studied. In this paper, we propose topology models to the Internet topology showing node addition and deletion.

In this work, we estimate the resilience of scale-free networks on a number of different attack methods. We study a number of different cases, where we assume that a small amount of knowledge on the network structure is available, or can be approximately estimated. We also present a class of real-life networks that prove to be very resilient on intentional attacks, or equivalently much more difficult to immunize completely than most model scale-free networks.

In this paper we propose an original modular extension of game theory named

games network

. The objective of games networks is to provide a theoretical framework which suits to modular dynamics resulting from different local interactions between various agents and which enables us to describe complex system in a modular way. Games networks describes situations where an agent can be involved in several different games, with several different other agents, at the same time. In particular, we focus on the determination of

global equilibria

, resulting from the composition of local equilibria for each game of the network.

However, several games networks can represent the same dynamics. We define the notion of dependence between agents, which allows us to compute a

games network normal form

. This normal form emphasizes the

elementary modules

which compose the games network.

We propose a new analysis method for a complex network based on a simple spin diffusion model. The model is constructed by a local interaction between vertices, as is in the spin dynamics. Several numerical experiments on network models are performed systematically under various initial conditions. The results strongly depend on the network structures, also on the initial conditions, while most conventional measures of the networks are almost same. It may be shown that the difference of dynamics induced by such interaction reveals a new characteristic feature of the network structure.

The analysis of statistical ensembles of networks by means of simulation is an important possibility to explore networks which emerge by optimization of some ’fitness’-function. In this paper, we compare the situations of the micro-, grand- and canonical ensemble based on their respective partition functions. We present results for a specific, recently introduced Hamiltonian. Interestingly, for all three ensembles we find scale-free networks with ’complex’ topology for a wide range of parameters. We further show results of some topological measures depending on energy and temperature.

In these notes we study synchronizability of dynamical processes defined on complex networks as well as its interplay with network topology. Building from a recent work by Barahona and Pecora [Phys. Rev. Lett.

89

, 054101 (2002)], we use a simulated annealing algorithm to construct optimally-synchronizable networks. The resulting structures, known as

entangled networks

, are characterized by an extremely homogeneous and interwoven topology: degree, distance, and betweenness distributions are all very narrow, with short average distances, large loops, and small modularity. Entangled networks exhibit an excellent (almost optimal) performance with respect to other flow or connectivity properties such as robustness, random walk minimal first-passage times, and good searchability. All this converts entangled networks in a powerful concept with optimal properties in many respects.

Content-based networks are introduced and their topological properties are outlined. A content-based model with Random Boolean dynamics, designed to mimic the gene regulation network, exhibits an increase in the number and complexity of attractors for increasing number of nodes. However, contrary to expectations based on Mean Field calculations for random scale-free networks, the attractors are not chaotic, even for average connectivities in excess of 2. Thus, the present model offers a promising tool for understanding complex biological networks.

In this paper we present a model where two interconnected network operators compete in linear prices in a market characterized by the existence of social connections among consumers, which are represented by a random regular graph. Assuming horizontal differentiation among operators, the customers select their network provider based on their preferences and the prices offered by the competing firms. In equilibrium the number of calls made to other agents depends on where they are located in the social network.

We present new algorithm for growth of non-clustered planar graphs by aggregation of cells with given distribution of size and constraint of connectivity

k

= 3 per node. The emergent graph structures are controlled by two parameters—chemical potential of the cell aggregation and the width of the cell size distribution. We compute several statistical properties of these graphs—fractal dimension of the perimeter, distribution of shortest paths between pairs of nodes and topological betweenness of nodes and links. We show how these topological properties depend on the control parameters of the aggregation process and discuss their relevance for the conduction of current in self-assembled nanopatterns.

We find the structural effect in geographical networks on the optimal paths and on the robustness of the connectivity. The communication efficiency are measured by the average path lengths and hops in the typical planar networks: Delaunay triangulation, random Apollonian network, and our proposed model with the well-balanced properties. The dynamic configuration will be useful especially for ad hoc communication.

We describe online collaborative communities by tripartite networks, the nodes being persons, items and tags. We introduce projection methods in order to uncover the structures of the networks, i.e. communities of users, genre families... The structuring of the network is visualised by using a tree representation. The notion of diversity in the system is also discussed.