Studies of dynamics of physical agent ecosystems

This thesis addresses the problem of learning in physical heterogeneous multi-agent systems (MAS) and the analysis of the benefits of using heterogeneous MAS with respect to homogeneous ones. An algorithm is developed for this task; building on a previous work on stability in distributed systems by Tad Hogg and Bernardo Huberman, and combining two phenomena observed in natural systems, task partition and hierarchical dominance. This algorithm is devised for allowing agents to learn which are the best tasks to perform on the basis of each agent's skills and the contribution to the team global performance. Agents learn by interacting with the environment and other teammates, and get rewards from the result of the actions they perform. This algorithm is specially designed for problems where all robots have to co-operate and work simultaneously towards the same goal. One example of such a problem is role distribution in a team of heterogeneous robots that form a soccer team, where all members take decisions and co-operate simultaneously. Soccer offers the possibility of conducting research in MAS, where co-operation plays a very important role in a dynamical and changing environment. For these reasons and the experience of the University of Girona in this domain, soccer has been selected as the test-bed for this research. In the case of soccer, tasks are grouped by means of roles. One of the most interesting features of this algorithm is that it endows MAS with a high adaptability to changes in the environment. It allows the team to perform their tasks, while adapting to the environment. This is studied in several cases, for changes in the environment and in the robot's body. Other features are also analysed, especially a parameter that defines the fitness (biological concept) of each agent in the system, which contributes to performance and team adaptability. The algorithm is applied later to allow agents to learn in teams of homogeneous and heterogeneous robots which roles they have to select, in order to maximise team performance. The teams are compared and the performance is evaluated in the games against three hand-coded teams and against the different homogeneous and heterogeneous teams built in this thesis. This section focuses on the analysis of performance and task partition, in order to study the benefits of heterogeneity in physical MAS. In order to study heterogeneity from a rigorous point of view, a diversity measure is developed building on the hierarchic social entropy defined by Tucker Balch. This is adapted to quantify physical diversity in robot teams. This tool presents very interesting features, as it can be used in the future to design heterogeneous teams on the basis of the knowledge on other teams.

Promoting resilience and protective factors in the Children's Fund: supporting children and young people's pathways towards social inclusion?

As part of the broader prevention and social inclusion agenda, concepts of risk, resilience, and protective factors inform a range of U.K. Government initiatives targeted towards children and young people in England, including Sure Start, the Children's Fund, On Track, and Connexions. This paper is based on findings from a large qualitative dataset of interviews conducted with children and their parents or caregiver who accessed Children's Fund services as part of National Evaluation of the Children's Fund research.1 Drawing on the notion of young people's trajectories, the paper discusses how Children's Fund services support children's and young people's pathways towards greater social inclusion. While many services help to build resilience and protective factors for individual children, the paper considers the extent to which services also promote resilience within the domains of the family, school, and wider community and, hence, attempt to tackle the complex, multi-dimensional aspects of social exclusion affecting children, young people, and their families.

Starting well: Gender, care and health in the family context

In this article we present a critique of a series of public policy documents that aim at improvement in health for the general population, particularly families, but fail to recognize or appreciate the implications of gender for the everyday and the long-term experiences of family members. Drawing upon considerations of gender, families, health time and space and previous theoretical work (McKie et al, 2002), we propose the concept of healthscapes to aid the analysis and development of public policies. A healthscapes approach allows analysis of health policy within the diverse and multi-dimensional notions of time, space and gender that infuse the lifecourse. We assert that consideration of the gendered and generational project of caring particularly in relation to the (re)production of health, should involve a reflective inter-play between theory research and policy.

Fast training of self organizing maps for the visual exploration of molecular compounds

Visual exploration of scientific data in life science area is a growing research field due to the large amount of available data. The Kohonen’s Self Organizing Map (SOM) is a widely used tool for visualization of multidimensional data. In this paper we present a fast learning algorithm for SOMs that uses a simulated annealing method to adapt the learning parameters. The algorithm has been adopted in a data analysis framework for the generation of similarity maps. Such maps provide an effective tool for the visual exploration of large and multi-dimensional input spaces. The approach has been applied to data generated during the High Throughput Screening of molecular compounds; the generated maps allow a visual exploration of molecules with similar topological properties. The experimental analysis on real world data from the National Cancer Institute shows the speed up of the proposed SOM training process in comparison to a traditional approach. The resulting visual landscape groups molecules with similar chemical properties in densely connected regions.

Scalability of efficient parallel K-Means

Clustering is defined as the grouping of similar items in a set, and is an important process within the field of data mining. As the amount of data for various applications continues to increase, in terms of its size and dimensionality, it is necessary to have efficient clustering methods. A popular clustering algorithm is K-Means, which adopts a greedy approach to produce a set of K-clusters with associated centres of mass, and uses a squared error distortion measure to determine convergence. Methods for improving the efficiency of K-Means have been largely explored in two main directions. The amount of computation can be significantly reduced by adopting a more efficient data structure, notably a multi-dimensional binary search tree (KD-Tree) to store either centroids or data points. A second direction is parallel processing, where data and computation loads are distributed over many processing nodes. However, little work has been done to provide a parallel formulation of the efficient sequential techniques based on KD-Trees. Such approaches are expected to have an irregular distribution of computation load and can suffer from load imbalance. This issue has so far limited the adoption of these efficient K-Means techniques in parallel computational environments. In this work, we provide a parallel formulation for the KD-Tree based K-Means algorithm and address its load balancing issues.

Process scheduling by output considerations

In multi-tasking systems when it is not possible to guarantee completion of all activities by specified times, the scheduling problem is not straightforward. Examples of this situation in real-time programming include the occurrence of alarm conditions and the buffering of output to peripherals in on-line facilities. The latter case is studied here with the hope of indicating one solution to the general problem.

KNIME: the Konstanz Information Miner

The Konstanz Information Miner is a modular environment which enables easy visual assembly and interactive execution of a data pipeline. It is designed as a teaching, research and collaboration platform, which enables easy integration of new algorithms, data manipulation or visualization methods as new modules or nodes. In this paper we describe some of the design aspects of the underlying architecture and briefly sketch how new nodes can be incorporated.

Predicting feed quality - chemical analysis and in vitro evaluation

As the ideal method of assessing the nutritive value of a feedstuff, namely offering it to the appropriate class of animal and recording the production response obtained, is neither practical nor cost effective a range of feed evaluation techniques have been developed. Each of these balances some degree of compromise with the practical situation against data generation. However, due to the impact of animal-feed interactions over and above that of feed composition, the target animal remains the ultimate arbitrator of nutritional value. In this review current in vitro feed evaluation techniques are examined according to the degree of animal-feed interaction. Chemical analysis provides absolute values and therefore differs from the majority of in vitro methods that simply rank feeds. However, with no host animal involvement, estimates of nutritional value are inferred by statistical association. In addition given the costs involved, the practical value of many analyses conducted should be reviewed. The in sacco technique has made a substantial contribution to both understanding rumen microbial degradative processes and the rapid evaluation of feeds, especially in developing countries. However, the numerous shortfalls of the technique, common to many in vitro methods, the desire to eliminate the use of surgically modified animals for routine feed evaluation, paralleled with improvements in in vitro techniques, will see this technique increasingly replaced. The majority of in vitro systems use substrate disappearance to assess degradation, however, this provides no information regarding the quantity of derived end-products available to the host animal. As measurement of volatile fatty acids or microbial biomass production greatly increases analytical costs, fermentation gas release, a simple and non-destructive measurement, has been used as an alternative. However, as gas release alone is of little use, gas-based systems, where both degradation and fermentation gas release are measured simultaneously, are attracting considerable interest. Alternative microbial inocula are being considered, as is the potential of using multi-enzyme systems to examine degradation dynamics. It is concluded that while chemical analysis will continue to form an indispensable part of feed evaluation, enhanced use will be made of increasingly complex in vitro systems. It is vital, however, the function and limitations of each methodology are fully understood and that the temptation to over-interpret the data is avoided so as to draw the appropriate conclusions. With careful selection and correct application in vitro systems offer powerful research tools with which to evaluate feedstuffs. (C) 2003 Elsevier B.V. All rights reserved.

Ecology of sponge assemblages (Porifera) in the Wakatobi region, south-east Sulawesi, Indonesia: richness and abundance

The influence of sedimentation, depth and substratum angle on sponge assemblages in the Wakatobi region, south-eastern Sulawesi, Indonesia was considered. Sponge assemblages were sampled from two reef localities. The first reef (Sampela) was highly impacted by high sedimentation rates with fine sediment particles that settle slowly, while the second (Hoga) experienced only fast settling coarse sediment with lower overall sedimentation rates. Sponge assemblages were sampled (area occupied and numbers) on the reef fiat (0 m) and at 5 (reef crest), 10 and 15 m (15 m at Hoga only). Some significant (P < 0.001) differences were observed in the area occupied and the number of sponge patches between surface angles and sites. Significantly lower (t > 4.61, df = 9, P < 0.001) sponge numbers, percentage cover and richness were associated with the reef flat at both sites compared with all other depths at each site, with the exception of abundance of sponges on the reef flat at Sampela, which was much greater than at any other depth sampled. Species richness increased with depth at both sites but differences between surface angles were only recorded at Sampela, with higher species richness being found on vertical, inclined and horizontal surfaces respectively A total of 100 sponge species (total area sampled 52.5 m(2)) was reported from the two sites, with 58 species found at Sampela and 71 species at Hoga (41% of species shared). Multi-dimensional scaling (MDS) indicated differences in assemblage structure between sites and most depth intervals, but not substratum angles. A number of biological (e.g. competition and predation) and physical (e.g. sedimentation and aerial exposure) factors were considered to control sponge abundance and richness. Unexpectedly a significant (F-1,F-169 = 148.98, P < 0.001) positive linear relationship was found between sponge density and area occupied. In areas of high sponge coverage, the number of patches was also high, possibly due to fragmentation of large sponges produced as a result of predation and physical disturbance. The MDS results were also the same whether sponge numbers or percentage cover estimates were used, suggesting that although these different approaches yield different sorts of information, the same assemblage structure can be identified.

Temporal and spatial variability of mobile fauna on a submarine cliff and boulder scree complex: a community in flux

The mobile component of a community inhabiting a submarine boulder scree/cliff was investigated at Lough Hyne, Ireland at dawn, midday, dusk and night over a 1-week period. Line transects (50 m) were placed in the infralittoral (6 m) and circumlittoral (18 m) zones and also the interface between these two zones (12 m). The dominant mobile fauna of this cliff consisted of echinoderms (6 species), crustaceans (10 species) and fish (23 species). A different component community was identified at each time/depth interval using Multi-Dimensional Scaling (MDS) even though both species diversity (Shannon-Wiener indices) and richness (number of species) remained constant. These changes in community composition provided indirect evidence for migration by these mobile organisms. However, little evidence was found for migration between different zones with the exception of the several wrasse species. These species were observed to spend the daytime foraging in the deeper zone, but returned to the upper zone at night presumably for protection from predators. For the majority of species, migration was considered to occur to cryptic habitats such as holes and crevices. The number of organisms declined during the night, although crustacean numbers peaked, while fish and echinoderms were most abundant during day, possibly due to predator-prey interactions. This submarine community is in a state of flux, whereby, community characteristics, including trophic and energetic relationships, varied over small temporal (daily) and spatial (m) scales.

Different supramolecular hydrogen bond structures and significant changes in magnetic properties in dinuclear mu(2)-1,1-N-3 copper(II) complexes with very similar tridentate Schiff base blocking ligands

Three new basal-apical, mu(2)-1,1-azide bridged complexes, [CuL1(N-3)](2) (1), [CuL2(N-3)](2) (2) and [CuL3(N-3)]2 (3) with very similar tridentate Schiff base blocking ligands [L-1=N-(3-aminopropyl) salicylaldimine, L-2=7-amino-4-methyl-5-azahept-3-en-2-one and L-3=8-amino-4-methyl-5-azaoct-3-en-2-one) have been synthesised and their molecular structures determined by X-ray crystallography. In complex 1, there is no inter-dimer H-bonding. However, complexes 2 and 3 form two different supramolecular structures in which the dinuclear entities are linked by strong H-bonds giving one-dimensional systems. Variable-temperature (300-2 K) magnetic susceptibility measurements and magnetization measurements at 2 K reveal that complexes 1 and 2 have antiferromagnetic coupling while 3 has ferromagnetic coupling which is also confirmed by EPR spectra at 4-300 K. Magnetostructural correlations have been made taking into consideration both the azido bridging ligands and the existence of intermolecular hydrogen bonds in complexes 2 and 3.

Supramolecular hydrogen-bond structures and magnetic interactions in basal-apical, dinuclear, azide-bridged copper(II) complexes

The synthesis, characterisation, X-ray single crystal structures and magnetic properties of three new basal-apical mu(2)-1,1-azide-bridged complexes [(CuLN3)-N-1](2) (1), [(CuLN3)-N-2](2) (2) and [(CuLN3)-N-3](2) (3) with very similar tridentate Schiff-base blocking ligands {HL1 = N-[2-(ethylamino) ethyl] salicylaldimine; HL2 = 7-(ethylamino)-4-methyl-5-azahept-3-en-2-one; HL3 = 7-amino-4-methyl-5-azaoct-3-en-2-one} have been reported [complex 1: monoclinic, P2(1)/c, a = 8.390(2), b = 7.512(2), c = 19.822(6) Angstrom, beta = 91.45(5)degrees; complex 2: monoclinic, P2(1)/c, a = 8.070(9), b = 9.787(12), c = 15.743(17) A, beta = 98.467(10)degrees; complex 3: monoclinic, P2(1)/n, a = 5.884(7), b = 16.147(18), c = 11.901(12) Angstrom, beta = 90.050(10)degrees]. The structures consist of neutral dinuclear entities resulting from the pairing of two mononuclear units through end-on azide bridges connecting an equatorial position of one copper centre to an axial position of the other, The copper ions adopt a (4+1) square-based geometry in all the complexes. In complex 2, there is no inter-dimer hydrogen-bonding. However, complexes 1 and 3 form two different supramolecular structures in which the dinuclear entities are linked by H-bonds giving one-dimensional systems. Variable temperature (300-2 K) magnetic-susceptibility measurements and magnetisation measurements at 2 K reveal that all three complexes have antiferromagnetic coupling. Magneto-structural correlations have been made taking into consideration both the azido bridging ligands and the existence of intermolecular hydrogen bonds. ((C) Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004).

Exact error estimates and optimal randomized algorithms for integration

Exact error estimates for evaluating multi-dimensional integrals are considered. An estimate is called exact if the rates of convergence for the low- and upper-bound estimate coincide. The algorithm with such an exact rate is called optimal. Such an algorithm has an unimprovable rate of convergence. The problem of existing exact estimates and optimal algorithms is discussed for some functional spaces that define the regularity of the integrand. Important for practical computations data classes are considered: classes of functions with bounded derivatives and Holder type conditions. The aim of the paper is to analyze the performance of two optimal classes of algorithms: deterministic and randomized for computing multidimensional integrals. It is also shown how the smoothness of the integrand can be exploited to construct better randomized algorithms.

Parallel Monte Carlo approach for integration of the rendering equation

This paper is addressed to the numerical solving of the rendering equation in realistic image creation. The rendering equation is integral equation describing the light propagation in a scene accordingly to a given illumination model. The used illumination model determines the kernel of the equation under consideration. Nowadays, widely used are the Monte Carlo methods for solving the rendering equation in order to create photorealistic images. In this work we consider the Monte Carlo solving of the rendering equation in the context of the parallel sampling scheme for hemisphere. Our aim is to apply this sampling scheme to stratified Monte Carlo integration method for parallel solving of the rendering equation. The domain for integration of the rendering equation is a hemisphere. We divide the hemispherical domain into a number of equal sub-domains of orthogonal spherical triangles. This domain partitioning allows to solve the rendering equation in parallel. It is known that the Neumann series represent the solution of the integral equation as a infinity sum of integrals. We approximate this sum with a desired truncation error (systematic error) receiving the fixed number of iteration. Then the rendering equation is solved iteratively using Monte Carlo approach. At each iteration we solve multi-dimensional integrals using uniform hemisphere partitioning scheme. An estimate of the rate of convergence is obtained using the stratified Monte Carlo method. This domain partitioning allows easy parallel realization and leads to convergence improvement of the Monte Carlo method. The high performance and Grid computing of the corresponding Monte Carlo scheme are discussed.

An autonomic approach in swarm-array computing

Can autonomic computing concepts be applied to traditional multi-core systems found in high performance computing environments? In this paper, we propose a novel synergy between parallel computing and swarm robotics to offer a new computing paradigm, `Swarm-Array Computing' that can harness and apply autonomic computing for parallel computing systems. One approach among three proposed approaches in swarm-array computing based on landscapes of intelligent cores, in which the cores of a parallel computing system are abstracted to swarm agents, is investigated. A task gets executed and transferred seamlessly between cores in the proposed approach thereby achieving self-ware properties that characterize autonomic computing. FPGAs are considered as an experimental platform taking into account its application in space robotics. The feasibility of the proposed approach is validated on the SeSAm multi-agent simulator.